AND8490/D Theory and Applications of the NCP1294, Switching Controller, and Associated Circuits for Lead Acid Battery Charging from a Solar Panel with Maximum Peak Power Tracking (MPPT) www.onsemi.com APPLICATION NOTE Introduction Background and System Requirements The following paper describes in detail the principle operation of the NCP1294, switching controller, and associated circuits for lead acid battery charging with MPPT. A design example and test data are included. The NCP1294 solar controller is a flexible solution used in Module Level Power Management (MLPM) solutions. Most of the solar controller battery chargers on the market today have a large digital content namely to manage maximum power point tracking and battery charging. The manipulation of the power stage, MPPT, and battery management can be controlled with analog circuitry in a cost effective manner. The solar controller parameters for the NCP1294 design are listed below in tabular form. Dusk Dawn Detection & Battery Discharge Management External LED Enable Solar Panel and Battery Enable Solar Panel Reverse Polarity Protection Reverse Polarity Protection Buck Boost Stage Battery Over Voltage MPP Circuit Battery Charge Circuit Under Voltage User Set Battery Voltage Remote Thermal Sensor User Set Charge Rate Figure 1. Block Diagram of ON Semiconductor NCP1294 120W Solar Controller © Semiconductor Components Industries, LLC, 2015 September, 2015 − Rev. 2 1 Publication Order Number: AND8490/D AND8490/D Table 1. ON SEMICONDUCTOR NCP1294 120 W SOLAR CONTROLLER SPECIFICATION SPECIFICATIONS Typical Units Output Current Rating 10 amp maximum (12 V battery controller is limited to 120 W) Nominal Battery Voltage 12/24/36 VDC PV Input Voltage 12 V to 60 VDC maximum (Recommended maximum VOC < 50 VDC) Power Consumption 0.30 W typical standby Charge Algorithm 3−stage Bulk/Acceptance/Float Absorption Voltage 14.3 VDC fixed (range 10.0 – 36.0 VDC) Float Voltage 13.5 VDC fixed (range 10.0 – 36.0 VDC) Voltage Set Point Limit 17.3 VDC fixed (range 10.0 – 36.0 VDC) Dusk Dawn Indicator Positive Logic Battery Low Indicator Positive Logic Temperature Compensation Optional sensor adjusts charge voltage based on battery temperature Power Conversion Efficiency 97% Typical 24 V to 14.3 V 8.39 A Output Physical Configuration and Dimensions 90 mm x 90 mm x 40 mm Figure 2. www.onsemi.com 2 AND8490/D design should also include safety features such as battery over voltage detection and solar panel under voltage detection. The resulting block diagram for the system is shown below. The specification described above will be configured in a block diagram (Figure 1) to aid the design process and ensure proper understanding of the system. The heart of the system is the power stage which must accept an input voltage of 12 V to 60 V and produce an output of 12 V to 36 V. Since the input voltage range spans the required output voltage, a buck boost topology is required to support the application. The topologies the designer could choose are SEPIC, Non Inverting buck boost, flyback, single switch forward, two switch forward, half bridge, full bridge, or others not listed. The output power for the converter is 120 W, thus the possibility of a flyback converter is eliminated. All of the topologies that are suitable for 120 W or greater have four or more switches and have the added complexity of a transformer with the exception of a four switch buck boost. Future work will include isolated topologies as power demands increase. The output voltage and battery charge rate should be selectable by the installing technician. The battery state of charge must be managed by the proper charge algorithm. Since the controller will be connected to a solar panel, it must have maximum power point tracking (MPPT) to provide a high value to the end customer. The controller should draw no more than 300 mW from either the output or input while not actively converting so all unnecessary circuitry should be turned off unless both the battery and the solar panel are installed. Two positive enable circuits have to be provided to external circuitry. One circuit detects the hours of darkness and another detects the state of charge of the battery so that the external circuit does not discharge the battery to the point of damage. The controller will be installed in the field by technicians and novices with varying degrees of experience, thus it is important to have reverse polarity protection for both the input and output. Since the controller and the batteries may be installed in either a hot or cold location, the controller must accommodate temperature compensation for charging the battery. The MPPT Methodology The power source for the controller specified is a solar panel. Solar panels have an IV curve that folds over shortly after the maximum power point where the output voltage falls as the current is increased beyond the maximum power point. Unfortunately the IV curves will change with irradiance, temperature, and age. Irradiance as defined by [3] is “the density of radiation incident on a given surface usually expressed in watts per square centimeter or square meter.” If the solar panel does not have mechanical sun tracking abilities, the irradiance will change as the sun moves approximately ±23° over a year. Further, the irradiance changes daily as the sun moves from horizon to horizon, resulting in a variation of power output throughout the day and year. The output power increases with decreasing temperature due in large part to the electron and hole mobility of the semiconductor material. Further, as temperature increases, the band gap energy of the semiconductor material also increases. Photons from the sun provide electrons in the valence band of the semiconductor with the energy to leap over the band gap into the semiconductor material. The larger band gap energy means more energy will be required from the photons in the sun to reach the conduction band. As a result, fewer electrons will reach the conduction band resulting in a less efficient solar cell. To extract the most power out of the variable source, the solar controller specified must employ MPPT. The MPPT must first locate the maximum power point and adjust for environmental conditions in a timely manner to keep the controller close to the maximum power point. Figure 3. IV Curve GeneCIS Solar Module 70W WSG0036E070 www.onsemi.com 3 AND8490/D Figure 4. Polycrystalline 6” Silicon Cells BP 3220T Temperature Curves [1] Figure 5. Polycrystalline 6” Silicon Cells BP 3220T Irradiance Curves [1] VMPP + k * Voc Many proven methods exist for MPPT for photo voltaic arrays, but only three will be discussed in this paper for brevity: Fractional Voc, Hill climber/Perturb and observe, and dynamic MPPT. (eq. 1) The k factor can change with temperature as shown in Figure 4, it is therefore beneficial to temperature compensate the fractional Voc circuit so that it is close to the maximum power point. The advantages to fractional Voc are that it is simple to implement and can be cost effective. The drawback is that the method is not good for a wide variety of solar panels with differing temperature coefficients and open circuit voltages. Fractional Open Circuit Voltage or Voc Fractional Open Circuit Voltage or Voc relies on the linear relationship between the open circuit voltage of the solar panel and the maximum power point. The method samples the voltage of the solar panel when no load is applied and holds it as a reference for the maximum power point voltage. Figure 5 shows that the maximum power point is between 71% and 78% [2] of the open circuit voltage or about 3/4. Therefore, the following equation can be used to calculate the MPP voltage: Hill Climber/Perturb and Observe Hill climber involves incrementing the duty cycle and calculating the resulting power change, where “perturb and observe” changes the voltage and observes the power www.onsemi.com 4 AND8490/D change. A PWM converter will be used to control the power extracted from the solar array which is a power limited source. Changing the duty cycle also changes the current draw from the array which in turn decreases the voltage in a power limited source. The relationship between duty cycle and voltage is linked so the two concepts will be combined and explained below. The converter starts by drawing very small power from the solar panel, the power is calculated, and the duty cycle is incremented and the power is again calculated. The previous power calculation is compared to the current power calculation and depending on the resulting increase or decrease; the duty cycle is incremented or decremented respectively as shown in Table 2. The process continues indefinitely finding the maximum power point as fast as the algorithm can move. One drawback to the method is that power is moved from its current point to another point which may not be the maximum power point to see if the maximum power point has changed. The method results in spending half of the time at the maximum power point, while half of the time remains at a suboptimal level. If the algorithm is long, this can mean a significant amount of time is spent not utilizing the maximum power of a solar panel. Dynamic MPPT Dynamic MPPT is applied as a change occurs in the system. Since the converter is using PWM, a change is occurring every switching cycle and power drawn from the solar panel is also changing in an observable way every cycle. Dynamic MPPT uses the voltage dip on the solar panel multiplied by the increasing current every switching cycle to determine the error signal that will be produced to regulate the duty cycle. The dynamic response detects the slope of the IV curve, creating a power ramp from which the error signal intersects creating a power representative duty cycle. The cycle ends when the ramp changes slope from positive to negative as shown in Figure 6. Table 2. HILL CLIMBER DECISION TABLE Change Change in Power Next Change Positive Positive Positive Positive Negative Negative Negative Positive Negative Negative Negative Positive Solar Panel Voltage The change in current is governed by the Inductance and the on time of the power stage Error Signal Error Signal Duty cycle is limited by the power ramp Solar Panel Current The ramp is a multiplication of the input current and input voltage. When the current and voltage are not at the MPPT point the slope is always increasing. Figure 6. Voltage and Current of a PWM Regulated Converter soft−start, accurate duty cycle limit control, less than 50 mA startup current, over and undervoltage protection, and bidirectional synchronization. [4] NCP1294 Part Description The NCP1294 fixed frequency feed forward voltage mode PWM controller contains all of the features necessary to be configured in a flyback, boost or forward topology and can be adapted for use in the buck, buck boost, half bridge, and full bridge topologies. The PWM controller has been optimized for high frequency primary side control operation. In addition, the device includes such features as: Switch Frequency and Maximum Duty Cycle Calculations The frequency of the NCP1294 can be set by using an oscillator timing capacitor, CT, which is charged by VREF www.onsemi.com 5 AND8490/D input voltage goes up, the rising edge of the ramp signal increases which reduces duty cycle to counteract the change. The feed forward feature can also be employed to provide a volt−second clamp, which limits the maximum product of input voltage and turn on time. The clamp is used in circuits, such as forward and flyback converter, to prevent the transformer from saturating. If the line voltage is much greater than the FF pin peak voltage, the charge current can be treated as a constant and is equal to VIN/R. Therefore, the volt−second value is determined by the following equation: through RT and discharged by an internal current source. During the discharge time, the internal clock signal sets the gate output to the low state, thus providing a user selectable maximum duty cycle clamp. Charge and discharge times are determined by following formulas; ƪ tc + RT * CT * ln ƪ td + RT * CT * ln tc td VPeak VValley = = = = V ref * V Valley V ref * V Peak ƫ V ref * V Peak * IdR T (eq. 2) ƫ V ref * V Valley * IdR T V IN * T ON + (VCOMP * VFF(d)) charging time discharging time peak voltage of the oscillator valley voltage of the oscillator VCOMP VFF(d) R C (eq. 3) = COMP pin voltage = FF pin discharge voltage As shown in the equation, the volt−second clamp is set by the VCOMP clamp voltage which is equal to 1.8 V. In forward or flyback circuits, the volt−second clamp value is designed to prevent transformers from saturation. In a buck or forward converter, volt−second is equal to: For the oscillator to function properly, RT has to be greater than 2.3k. Feed Forward Voltage Mode Control In conventional voltage mode control, the ramp signal has a fixed rising and falling slope. The feedback signal is derived solely from the output voltage. Consequently, voltage mode control has inferior line regulation and audio susceptibility. Feed forward voltage mode control derives the ramp signal from the input line. Therefore, the ramp of the slope varies with the input voltage. At the start of each switch cycle, the capacitor connected to the FF pin is charged through a resistor connected to the input voltage. Meanwhile, the gate output is turned on to drive an external power switching device. When the FF pin voltage reaches the error amplifier output VCOMP, the PWM comparator turns off the gate, which in turn opens the external switch. Simultaneously, the FF capacitor is quickly discharged to 0.3 V. Overall, the dynamics of the duty cycle are controlled by both input and output voltages. For example, an elevated output voltage reduces VCOMP which in turn causes duty cycle to decrease. However, if the input voltage varies, the slope of the ramp signal will react immediately which provides a much improved line transient response. When the V IN * T ON + n V OUT Ts (eq. 4) n = transformer turns ratio The constant, n, is determined by the regulated output voltage, switching period, and transformer turns ratio (use 1.0 for buck converter). As shown in Equations 3 and 4 during steady state, VCOMP doesn’t change for input voltage variations, intuitively explaining why FF voltage mode control has superior line regulation and line transient response. Knowing the nominal value of VIN and TON, one can also select the value of RC to place VCOMP at the center of its dynamic range. Design Procedure When selecting a topology for the solar controller, it is important to understand the converters basic operation and its limitations. The topology selected is the non−inverting four switch non−synchronous buck boost topology. The converter operates with a single control signal from the NCP1294, which turns on Q1 and Q2 simultaneously charging L1. L1 Q1 D2 VSW2 VSW1 D1 Q2 Drive Figure 7. www.onsemi.com 6 AND8490/D Drive Drive VSW 1 VSW 1 VSW 2 VSW 2 Inductor Current Inductor Current IL2 IL1 IL1 Q1 Q1 Q2 Q2 D1 D1 D2 D2 Figure 8. Buck Boost Operation Time Delayed Left, Ideal Right DBUCK_BOOST FSW T TOFF TON VIN VOUT The four switch buck boost topology is show in Figure 8 where a single inductor is used to control the voltage and current. The four switch non−inverting buck boost has two modes of operation that will be considered the buck mode and the buck boost mode. In the buck mode, the converter produces input voltage pulses that are LC filtered to produce a lower DC output voltage VOUT. The output voltage can be changed by modifying the on time relative to the switching period T or switching frequency. The ratio of high side switch on time to the switching period is called duty ratio D. F sw + D+ D BUCK + D BUCK_BOOST + T ON T 1 T (1 * D ) + V OUT V IN ³ 56.3% + V OUT V OUT ) V IN (eq. 6) T 13.5 V 24 V ³ 36% + 13.5 V 13.5 V ) 24 V (eq. 7) D DBUCK Buck boost converter duty cycle Switching frequency Switching period High side switch off time High side switch on time Input voltage Output voltage The solar controller will operate in buck mode if the output voltage can be achieved from 1% to 89%. If the output voltage cannot be reached due to duty cycle limitations, it will switch to buck boost mode where the voltage can be achieved. The duty cycle will change from 89% to a lower duty cycle as show in Figure 9. One thing to note is that when the converters mode switches from buck to buck boost, the error signal will take time to change the duty cycle. The instantaneous change of mode will leave a buck boost converter trying to switch at 89% duty cycle and trying to slew to 47%; this results in the converter trying to output 130 V at the trade over region. The NCP1294 is equipped with a pulse by pulse current limiter and will stop the converter from building energy in the choke to dangerous proportions, easing the transition in duty cycle. (eq. 5) T OFF = = = = = = = = Duty cycle = Buck converter duty cycle www.onsemi.com 7 AND8490/D Figure 9. Transfer Ratio Between Buck and Boost Mode for Multiple Batteries Inductor Selection 1*D 11.06 mH + (eq. 8) D = DBUCK = DBUCK_Boost = ratio FSW = IOUT = = LOUT_BUCK LOUT_BUCK_BOOST = ra = = Ripple current = Output current = Ripple current ratio Using the ripple current rule of thumb, the user can establish acceptable values of inductance for a design using Equation 9. L OUT_BUCK + 26.88 mH + V IN * D BUCK I OUT ra F SW ³ (eq. 9) 24 V * 56% 10 A 25% I OUT 200 kHz www.onsemi.com 8 ra BUCK_Boost ³ F SW 24 * 36% 10 A 1*36% DI ra + I out DI Iout ra V IN * D BUCK_Boost L OUT_BUCK_BOOST + When selecting an inductor, the designer may employ a rule of thumb for the design where the percentage of ripple current in the inductor should be between 10% and 40%. The ratio of ripple current to maximum output current is given in Equation 8. 25% 200 kHz Duty ratio Buck converter duty cycle ratio Buck boost converter duty cycle Switching frequency Output current Buck output inductance Buck boost output inductance Ripple current ratio AND8490/D 3 Batteries 2 Batteries 15 μH 1 Battery Figure 10. Inductance vs. Current Ripple Ratio If an inductance of 22 mH is chosen and the converter charges 1 to 3 batteries, the ripple current requirement is satisfied in the buck boost mode; however in the buck mode, the ripple requirement is only met for the case of one battery. The chart below shows the resulting ripple current from the selection of 22 mH. Figure 11. Calculated Ripple Current Ratio with 22 H When selecting an inductor, the designer must not exceed the current rating of the part. To keep within the bounds of the part’s maximum rating, a calculation of the RMS current and peak current are required. I DC_Buck + I OUT 10.011 A + 10 A Ǹ1 ra 2 ³ 12 Ǹ1 ) 30% 12 I DC_Buck_Boost + 15.63 A + DBUCK_Boost IOUT IDC_Buck (eq. 10) 2 www.onsemi.com 9 I OUT 1 * D BUCK_Boost ³ 10 A 1 * 36% = Buck boost converter duty cycle ratio = Output current = Buck converter inductor RMS current AND8490/D IDC_Buck _Boost ra = Buck boost converter inductor RMS current = Ripple current ratio I PK_BUCK + I OUT 11.55 A + 10 A I PK_BUCK_Boost + 18.05 A + DBUCK_Boost IOUT IPK_BUCK IPK_BUCK_Boost ǒ1 ) raǓ ³ 2 ra (eq. 11) ǒ1 ) 31%Ǔ = = = = Buck boost converter duty cycle ratio Output current Buck converter inductor peak current Buck boost converter inductor peak current = Ripple current ratio 2 I OUT 1 * D BUCK_Boost 10 1 * 36% ǒ1 ) raǓ ³ 2 ǒ1 ) 31%Ǔ 2 Figure 12. Calculated DC Current in Inductor for Buck Mode and Buck Boost Mode Figure 13. Calculated Peak Current in the Inductor for Buck and Buck Boost Mode www.onsemi.com 10 AND8490/D LP tot + LP CU_DC ) LP CU_AC ) LP Core ³ A standard inductor should be found so the inductor will be rounded to 15 mH. The inductor should support an RMS current of 20.5 A and a peak current of 22 A. A good design practice is to select an inductor that has a saturation current that exceeds the maximum current limit with some margin. The final selection of an output inductor has both mechanical and electrical considerations. From a mechanical perspective, smaller inductor values generally correspond to smaller physical size. Since the inductor is often one of the largest components in the regulation system, a minimum inductor value is particularly important in space constrained applications. Consequently, output capacitors must supply the load current until the inductor current reaches the output load current level. The peak−to−peak ripple current for NCP1294 is given by the following equation: I pp + V OUT L OUT (1 * D ) F SW D FSW Ipp LOUT VOUT = = = = = ³ 1.02 A + 13.5 V 22 mH 285 mW + 260 mW ) 1 mW ) 24 mW LPCore LPCU_AC LPCU_DC LPtot The important factors to consider when selecting an output capacitor are DC voltage rating, ripple current rating, output ripple voltage requirements, and transient response requirements. The output capacitor must be able to operate properly for the lifetime of a product. When selecting a capacitor, it is important to select a voltage rating that is derated to the guaranteed operating lifetime of a product. Further, it is important to note that when using ceramic capacitors, the capacitance decreases as the voltage applied increases; thus a ceramic capacitor rated at 22 mF 25 V may measure 4.4 mF with an applied voltage of 13.5 V depending on the type of capacitor selected. The output capacitor must be rated to handle the ripple current at full load with proper derating. Since the load is a battery, the question may arise, isn’t a battery just like a capacitor? The lead acid battery which is the load is modeled as a voltage source in series with an RC network as shown in Figure 14. (1 * 56%) 200 kHz (eq. 12) From Equation 12, it is clear that the ripple current increases as LOUT decreases, emphasizing the trade−off between dynamic response and ripple current. The power dissipation of an inductor falls into two categories: copper and core losses. Copper losses can be further categorized into DC losses and AC losses. A good first order approximation of the inductor losses can be made using the DC resistance as shown below: 260 mW + 10 2 DCR ³ 2 635.2 mW + 15.63 2 DCR IDC_BUCK IDC_BUCK Boost LPCU_DCB LPCU_DCBB CF = 1nF R 10mW Ro = 20mW (eq. 13) 2.6 mW LP CU_DCB + I DC_BUCK_Boost 2 Inductor core power dissipation Inductor AC power dissipation Inductor DC power dissipation Total inductor losses Output Capacitor Selection Duty ratio Switching frequency Peak−to−peak current of the inductor Output inductance Output voltage LP CU_DCB + I DC_BUCK = = = = (eq. 14) DCR ³ 2.6 mW Figure 14. Thevenin Equivalent Battery Model for Lead Acid Battery = Inductor DC resistance = Buck regulator inductor DC current = Buck boost regulator inductor DC current = Buck inductor DC power dissipation = Buck boost inductor DC power dissipation The series resistance is very low to begin with; therefore the majority of the pulsed current will go to the battery. The only time the capacitor is needed is when the current flow from the battery decreases towards the end of the absorption stage and in the float stage. Thus a small maintaining capacitance is needed, but ripple voltage and noise are of small concern for this type of battery. The capacitor RMS ratings given in datasheets are generally for lower switching frequency than used in switch mode power supplies, but a multiplier is given for higher The core losses and AC copper losses will depend on the geometry of the selected core, core material, and wire used. Most vendors will provide the appropriate information to make accurate calculations of the power dissipation at which point the total inductor losses can be captured by the equation below: www.onsemi.com 11 AND8490/D frequency operation. The RMS current for the output capacitor can be calculated below: CO RMS_BUCK + I OUT CINESR IinRMS PCIN ra 31% ³ 0.294 A + 10 A Ǹ12 Ǹ12 (eq. 15) Ǹ CO RMS−BUCK−Boost + I OUT 11.54 A + 10 CoRMS_BUCK CoRMS_BUCK_Boost DBUCK_Boost IOUT ra Ǹ Due to large di/dt through the input capacitors, electrolytic or ceramics should be used. If a tantalum capacitor must be used, it must be surge protected, otherwise capacitor failure could occur. 2 D BUCK_Boost ) ra12 1 * D BUCK_Boost 56% ) 56% 12 ³ Power MOSFET Dissipation 2 Power dissipation, package size, and the thermal environment drive power supply design. Once the dissipation is known, the thermal impedance can be calculated to prevent the specified maximum junction temperatures from being exceeded at the highest ambient temperature. Power dissipation has two primary contributors: conduction losses and switching losses. The high−side MOSFET will display both switching and conduction losses. The switching losses of the low side MOSFET will not be calculated as it switches into nearly zero voltage and the losses are insignificant. However, the body diode in the low−side MOSFET will suffer diode losses during the non−overlap time of the gate drivers. Starting with the high−side buck and boost MOSFET, the power dissipation can be approximated from the following equation: 1 * 56% = Buck converter output capacitor RMS current = Buck boost converter output capacitor RMS current = Buck boost converter duty cycle ratio = Output current = Ripple current ratio Input Capacitor Selection The input capacitor has to sustain the ripple current produced during the on time of the upper MOSFET, so it must have a low ESR to minimize losses and input voltage ripple. The RMS value of the input ripple current is: ǸDBuck Iin RMS_Buck + I OUT 1.34 A + 4.96 A Iin RMS_Buck_Boost + Ǹ P D_HS + P COND ) P SW_TOT PCOND PD_HS (eq. 16) (1 * 56%) PSW_TOT I OUT 1 * D Buck_Boost D Buck_Boost 11.61 A + ǒ1 * D BUCKǓ ³ Ǹ56% 10 1 * 56% ƪ1 * D Buck_Boost ) Ǹ 56% ƫ ra 2 ³ 12 ƪ1 * 56% ) 56% ƫ 12 I RMS_BUCK_Boost + 2 DBUCK = DBUCK_Boost = ra = IOUT = IRMS_BUCK = IRMS_BUCK_Boost = 246 mW + 10 mW ǒ4.96 AǓ 2 I OUT 1*D Ǹ ǒ1 ) ra12 Ǔ 2 D Ǹ D (eq. 19) ǒ1 ) ra12 Ǔ 2 Buck duty ratio Buck boost duty ratio Ripple current ratio Output current Buck high side MOSFET RMS current Buck boost high side MOSFET RMS current The first term in Equation 18 is the conduction loss of the high−side MOSFET while it is on. Since the low side MOSFET is on at the same time as the boost MOSFET the conduction loss can be approximated the same. The equation reaches its maximum value with D = 0.5 at which point the input capacitance RMS current is half the output current. Loss in the input capacitors can be calculated with the following equation: ǒIin RMSǓ = Conduction losses = Power losses in the high side MOSFET = Total switching losses I RMS_BUCK + I OUT Duty ratio Buck regulator duty cycle ratio Buck boost regulator duty cycle ratio Buck input capacitance RMS current Buck boost input capacitance RMS current = Load current P CIN + CIN ESR (eq. 18) Using the ra term from Equation 8, IRMS becomes: D = DBuck = DBuck_Boost = IinRMS_Buck = IinRMS_Buck_Boost= IOUT = Input capacitance Equivalent Series Resistance = Input capacitance RMS current = Power loss in the input capacitor ³ (eq. 17) 2 www.onsemi.com 12 AND8490/D P COND_BUCK + ǒI RMS_HS_BUCKǓ 2 R DS(on) Vth (eq. 20) P RMS_BUCK_Boost + ǒI RMS_HS_BUCK_BoostǓ 2 R DS(on) IRMS_HS_BUCK = Buck RMS current in the high side MOSFET IRMS_HS_BUCK_Boost = Buck boost RMS current in the high side MOSFET PCOND_BUCK = Buck conduction power losses PCOND_BUCK_Boost = Buck boost conduction power losses RDS(ON) = On resistance of the high side MOSFET The second term from Equation 18 is the total switching loss and can be approximated from the following equations. P SW_TOT + P SW ) P DS ) P RR PDS Figure 15. High Side MOSFET Total Charge (eq. 21) t RISE + = High side MOSFET drain to source losses = High side MOSFET reverse recovery losses = High side MOSFET switching losses = High side MOSFET total switching losses PRR PSW PSW_TOT ǒI OUT (eq. 22) P SW_BUCK_Boost + P TON ) P TOFF + ǒ I OUT 1*D FSW IOUT PSW_BUCK PSW_BUCK_Boost PTON PTOFF tFALL tRISE VIN V IN Ǔ F SW t FALL + IG12 1 2 QGD RG RHSPD tFALL VBST VTH ǒtRISE ) t FALLǓ = Switching frequency = Load current = Buck high side MOSFET switching losses = Buck boost high side MOSFET switching losses = Turn on power losses = Turn off power losses = MOSFET fall time = MOSFET rise time = Input voltage Q GD ǒV BST * V THǓńǒR HSPU ) R GǓ (eq. 23) = Output current from the high−side gate drive = MOSFET gate to drain gate charge = Drive pull up resistance = MOSFET gate resistance = MOSFET rise time = Boost voltage = MOSFET gate threshold voltage QGD RHSPU RG tRISE VBST VTH F SWǓ V IN I G1 + IG1 The first term for total switching losses from Equation 21 are the losses associated with turning the high−side MOSFET on and off and the corresponding overlap in drain voltage and current. 1 P SW_BUCK + P TON ) P TOFF + 2 ǒt RISE ) tFALLǓ Q GD Q GD I G2 + Q GD ǒVBST * V THǓńǒR HSPD ) R GǓ (eq. 24) = Output current from the low−side gate drive = MOSFET gate to drain gate charge = MOSFET gate resistance = Drive pull down resistance = MOSFET fall time = Boost voltage = MOSFET gate threshold voltage Next, the MOSFET output capacitance losses are caused by both the high−side and low−side MOSFETs, but are dissipated only in the high−side MOSFET. P DS + COSS FSW PDS When calculating the rise time and fall time of the high side MOSFET it is important to know the charge characteristic shown in Figure 15. VIN www.onsemi.com 13 1 @ C OSS 2 V IN 2 F SW (eq. 25) = MOSFET output capacitance at 0 V = Switching frequency = MOSFET drain to source charge losses = Input voltage AND8490/D Finally, the loss due to the reverse recovery time of the body diode in the low−side MOSFET is calculated as follows: P RR + Q RR @ V IN @ F SW FSW PRR NOLLH (eq. 26) = Switching frequency = High side MOSFET reverse recovery losses = Reverse recovery charge = Input voltage QRR VIN PBODY VFD Compensation Network To create a stable power supply, the compensation network around the error amplifier must be used in conjunction with the PWM generator and the power stage. Since the power stage design criteria is set by the application, the compensation network must correct the overall output to ensure stability. The NCP1294 is a voltage mode voltage feed forward part and as such there exists a voltage loop with an input voltage modified ramp. The output inductor and capacitor of the power stage form a double pole and the loop must compensate for that. The ESR of the output capacitor creates a “zero” at the frequency as shown in Equation 31: The NCP1294 demonstration board is setup to use either a non synchronous or synchronous buck stage depending on the user’s needs. If a synchronous buck is required, the calculations and explanation for the dissipation are as follows. The low−side MOSFET turns on into small negative voltages so switching losses are negligible. The low−side MOSFET’s power dissipation only consists of conduction loss due to RDS(on) and body diode loss during non−overlap periods. If a non synchronous design is used P conduction goes to zero and PBODY becomes PDIODE. P D_LS + P COND ) P BODY PBODY PCOND PD_LS (eq. 27) FZ ESR + = Low side MOSFET body diode losses = Low side MOSFET conduction losses = Low side MOSFET losses IRMS_LS RDS(ON)_LS PCOND (eq. 28) R DS(on)_LS = = = = Ǹǒ 1 * D BUCKǓ ǒ1 ) ra12 Ǔ 2 F SW (eq. 29) 2p 640 mF 0.010 mW 1 ǸLOUT 2p ³ Ǹ22 mH 2p FZ P_BUCK + 2p = Switching frequency = Load current = Dead time between the high−side www.onsemi.com 14 ǸLOUT 2p 1 * 47% COUT DBUCK_Boost FP_BUCK FP_BUCK_Boost LOUT (eq. 32) 640 mF 1 * D BUCK_Boost 590 Hz + ǒNOLLH ) NOLHLǓ ³ C OUT 1 1.34 kHz + (eq. 30) FSW IOUT NOLHL (eq. 31) = Output capacitor ESR = Output capacitor = Output capacitor zero ESR frequency FZ P_BUCK + Buck duty ratio Load current RMS current in the low side Ripple current ratio I OUT ³ The frequency of the double pole varies for both a buck and a buck boost, they can be calculated as shown below: The body diode losses can be approximated as: P BODY + V FD C OUT 1 COESR COUT FZESR = RMS current in the low side = Low−side MOSFET on resistance = High side MOSFET conduction losses I RMS_LS + I OUT @ DBUCK IOUT IRMS_LS ra 2 1 CO ESR 2p 24.87 kHz + Conduction loss in the low−side MOSFET is calculated as follows: P COND + ǒI RMS_LSǓ MOSFET turning off and the low−side MOSFET turning on = Dead time between the low−side MOSFET turning off and the high−side MOSFET turning on = Low−side MOSFET body diode losses = Body diode forward voltage drop Ǹ22 mH = = = = = ³ ³ C OUT 640 mF ³ Output capacitor Buck boost duty ratio Buck double pole frequency Buck boost double pole frequency Output inductance AND8490/D Figure 16. Pole Frequency vs. Input Voltage and Mode The key is to compensating a combination buck and buck boost converter is to know where the RHPZ is and how it changes with load as is defined by the equation above. The RHPZ is at its lowest point when the load is the highest, for this application that occurs with one battery and when the output voltage is at its lowest, as it is a squared term. The input voltage at its highest voltage will also yield the lowest RHPZ as shown in the figure below. Note that when the converter switches to buck mode, RHPZ does not exist, thus in the figure, buck mode is shown as 1 MHz RHPZ although it is understood that it does not exist. The right half plane zero for the buck boost converter can be calculated with the following equation. FZ RHPZ_BUCK_Boost + 12.5 kHz + V IN 2 2p ǒVOUT ) VINǓ ³ L OUT (eq. 33) 24 V 2 2p ǒ13.5 V ) 24Ǔ FRHPZ_BUCK_Boost IOUT LOUT VIN VOUT = = = = = I OUT 22 mH 8.89 A Right half plane zero Output current Output inductance Input Voltage Output Voltage Figure 17. Right Half Plane Zero Frequency www.onsemi.com 15 AND8490/D The three equations above define the bode plot that the power stage has created or open loop response of the system. The next step is to close the loop by considering the feedback values. The closed loop crossover frequency should be less than 1/10 of the switching frequency. The designer must balance a high crossover frequency with crossing over a decade before the RHPZ, thus the designer must determine which frequency is lower. In this application the lowest RHPZ is at 6.721 kHz, therefore the crossover frequency will be set to 1.5 kHz. Figure 18 shows type III error amplifier compensation to make the loop stable. Only the type III compensation is shown as type II and type I are a subset. CP The relationship between the resistor divider network above and the output voltage is shown in Equation 34: (eq. 34) V OUT * V REF = = = = Top resistor divider Bottom resistor divider Output voltage Regulator reference voltage The most frequently used output voltages and their associated standard R1 and R2 values are listed in Table 3. Table 3. OUTPUT VOLTAGE SETTINGS VOUT RC CC Ǔ V REF R1 R2 VOUT VREF VOUT(V) R1(kW) R2(kW) 13.5 100 10.2 27 100 4.87 40.5 100 3.24 CF R1 RF The compensation components for the Type III error amplifier can be calculated using the method described below. The method serves to provide a good starting place for compensation of a power supply. The values can be adjusted in real time using the compensation tool CompCalc for the buck portion of the converter but must be checked in the application http://www.onsemi.com/pub/Collateral/COMPCALC.ZIP The integrator for the operation amplifier is set as follows. The designer should set the voltage feed forward resistor to the same value as R1 so that the two values cancel out: FB COMP ǒ R2 + R1 EA R2 VREF Figure 18. Type III Error Amplifier Configuration The compensation network consists of the internal error amplifier and the impedance networks. The compensation network has to provide a closed loop transfer function with the highest 0 dB crossing frequency to have fast response and the highest gain in DC conditions to minimize load regulation issues. A stable control loop has a gain crossing with −20 dB/decade slope and a phase margin greater than 45°. Include worst−case component variations when determining phase margin. To start the design, a resistor value should be chosen for R1 from which all other components can be chosen. A good starting value is 100 kW. The NCP1294 allows the output of the DC−DC regulator to be adjusted down to 1.263 V via an external resistor divider network. The regulator will maintain 1.263 V at the feedback pin. Thus, if a resistor divider circuit was placed across the feedback pin to VOUT, the regulator will regulate the output voltage proportional to the resistor divider network in order to maintain 1.263 V at the FB pin. C1 + V IN 2p * Vramp * Fcross ³ V IN * R * C * F sw 2p * V IN * D * Fcross * R1 ³ (eq. 35) R * C * F sw 2p * D * Fcross * R1 41.2 nF + C C1 D FSW Fcross R R1 VIN Vramp 100 kW * 1.5 nF * 200 kHz 2p * 89% * 1.3 kHz * 100 kW = = = = = = = = = Feed forward capacitor Compensation capacitor Maximum duty cycle of the converter Switching frequency Crossover frequency Feed forward resistor Upper resistor divider Input Voltage Ramp Voltage The pole frequency for C1 and R1 should now be calculated. FPO + 38.56 Hz + Figure 19. Feedback Resistor Divider www.onsemi.com 16 1 2p * C1 * R1 1 2p * 41.26 nF * 100 kW (eq. 36) AND8490/D C1 FPO R1 System Turn On and Battery Current Draw = Compensation capacitor = Pole frequency = Upper resistor divider The system being created is connected to two finite sources that will supply power to the load at different times of the day and will more than likely not work at the same time except for brief periods. The system is not complete without both the battery and the solar panel installed therefore, detection of the presence of the battery load and the solar panel source are advantageous. For instance it would not be good to dissipate solar panel energy in providing battery voltage if no battery is connected. If a solar panel is connected, the battery would be drained looking for a solar panel to be connected. A simple solution for checking that both the solar panel is connected and the battery is connected is a low current draw comparator. Q1 serves as the comparator and checks to determine if there is a voltage greater than the base−emitter saturation voltage. If the resistor divided solar panel voltage is greater than the VBE(sat) of the NPN transistor, then Q1 becomes saturated and opens Q2, allowing voltage into power the other circuits. If a battery is not provided, none of the circuitry turns on. If a battery is detected and no solar panel is installed, Q2 remains off as does the circuitry. The solar panel detection circuit should be set low so it does not interfere with the operation of the MPPT. The zero should be placed at the double pole of the plant. ǒ 2.175 nF + 1 * ³ 1 24.87 kHz (eq. 37) Ǔ = Compensation capacitor = Double pole frequency of buck boost converter = ESR frequency = Upper resistor divider FZESR R1 FPO F P_BUCK_Boost FP_BUCK ³ 5.425 kW + 100 kW 38.567 Hz 710.9 Hz (eq. 38) = Double pole frequency of buck converter = Double pole frequency of buck boost converter FP_BUCK_Boost FPO R1 R2 C3 + ǒ 1 2p * 100 kW 20 Hz C2 FP_BUCK_Boost R2 + R1 Ǔ 1 1 1 * 4pR1 FP _BUCK FZ ESR C2 + = Pole frequency of R1 and C1 = Upper resistor divider = Resistor 2p Solar Panel Battery Voltage Voltage Q2 Circuit Power 1 (R2 * Fcross * R1 * FPO) 49.796 nF + 2p C3 Fcross FPO R1 R2 ON (eq. 39) 1 ǒ5.425 kW * 1.3 kHz * 100 kW * 38.567 HzǓ = = = = = Q1 Compensation pole capacitor Crossover frequency Pole frequency of R1 and C1 Upper resistor divider Resistor Figure 20. Dual Solar Panel Battery Detection Circuit If the ESR frequency is greater than the switching frequency, a CF compensation capacitor may be needed for stability as the output LC filter is considered high Q and thus will not give the phase boost at the crossover frequency. R3 + 2.943 kW + R1 * F BUCK_Boost FZ ESR * F P_BUCK_Boost 100 kW * 710 Hz ³ (eq. 40) 24.87 kHz * 710 Hz CP Fcross FP_BUCK_Boost FZESR The circuit power comes directly from the output battery and the intention of system is to be easily adaptable to a single battery 12 V, dual battery 24 V, or triple battery 36 V system. Therefore it is necessary to provide a dependable voltage to the measurement and switching circuitry. In this application the MC78L08 is used and will provide a steady 8 V supply while handling as much as 40 V. = Compensation pole capacitor = Cross over frequency = Double pole frequency of buck boost converter = ESR frequency www.onsemi.com 17 AND8490/D voltage is greater than 2.5 V, D2 pulls down on R3 and Q1 is turned off, pulling up on the LED enable signal. Further, the circuit in the figure below is linked to the battery and solar panel detection circuit. The ON node connection detects daylight and disables the LED string to prevent operation of the LEDs while the solar panel is charging the batteries. Regulated Circuit Votlage Circuit Power LDO Battery Voltage Figure 21. LDO Circuit The system charges batteries during the day time and discharges them during the night to illuminate a defined space. The input energy is not guaranteed, but the output energy remains constant over a long period of time. If a system is not sized properly, the discharge of the battery can result in damage. To stop the battery from being damaged, the LED circuit must be inhibited from operation if the battery is depleted. The circuit below uses a TL431 to monitor the battery voltage and provides an enable disable signal to the LEDs when the battery is below the cutoff voltage. If the R1 and R2 battery divided voltage is less than 2.5 V, D1 does not pull down on R3. R3 provides greater than the MOSFET threshold voltage to Q1 giving the LED a disable signal. Conversely, if the R1 and R2 battery divided R6 R1 D2 R3 ON R4 Q1 D1 R5 R2 LED Enable/Disable Signal Figure 22. Battery Discharge and Daylight Detection Circuit www.onsemi.com 18 AND8490/D Input and Output Current Measurements and out of the charger. The simplest method of current sense is low side sensing where the current running in the return path is sensed by an amplifier as shown in Figure 23. To implement battery charging and MPPT it is important to sense the input output currents to control the power into Battery Drive Solar Panel Current Sense Resistor Return Current − Amplified Voltage + Sense Amplifier Figure 23. Low Side Amplifier Current Sensing The issue with low side sensing is that the solar controller being produced may be configured in a system with many other solar controllers and the installer may want to employ central grounding which will defeat the current sense employed by diverting the current from the sense resistor in part or in full. The current sensed by the amplifier then becomes small in comparison with the actual current being drawn from the device. While low side current sense is easy to implement and can be executed inexpensively, it may not provide a positive user experience as the controller will cease to function properly when installed by different users. Figure 24 shows central grounding and how it can affect the sensed current in low side sensing. Battery Drive Solar Panel Current Sense Resistor 10mW Return Current 1 Return Current 2 Central Grounding Point1mW Figure 24. Central Grounding Diagram Implementing high side current sensing allows common point grounding as the ground path current does not affect this sensing. Unfortunately the operation amplifier chosen must have a wide common mode input range. The amplifier chosen must work at high input voltages and the complexity of the circuit increases. www.onsemi.com 19 AND8490/D VREF Sense Amplifiers R2 = 10k + Amplified Voltage − R4 = 10k R1 = 1k R3 = 1k Current Sense Resistor Battery Drive Solar Panel Figure 25. Amplifier High Side Current Sensing multiple of 2 of the base current. The next current mirror considered is the Wilson current mirror, which has relatively good current parity. The only problem with the Wilson current mirror is that the number of transistors doubles from 4 to 8, making it unattractive from a cost and circuit management perspective. Since the primary aim of the system is making a cost effective solution, the designer should aim to implement the solution minimizing current consumption and maximizing accuracy. One cost effective way to sense high side current is to use current mirrors to mirror the voltage to ground level. The simplest current mirror to use is a Windlar, which is handicapped by the Early effect and the current differs by a B Current Flow IN A Current Flow IN R1= 499 W R2= 4.9 kW R1= 499 W IOUT IOUT Current Sense Resistor Current Sense Resistor R3= 4.9 kW R2= 4.9 kW R3= 4.9 kW Figure 26. Left: Windlar High Side Current Sense, Right: Wilson High Current Sense www.onsemi.com 20 AND8490/D Figure 27. Simulated Accuracy of the Windlar Current Sense Figure 28. Simulated Accuracy of the Wilson Current Sense www.onsemi.com 21 AND8490/D Figure 29. Temperature Error of Windlar and Wilson Current Senses With the error introduced by the Windlar current sense, it may first appear that this current sense is not useful, but if a relative current level is desired, the circuit may still be useful. For most MPPT algorithms the maxima of a voltage multiplied by a current is desired. If the current measurement changes with voltage, the maximum is found but perhaps not the optimal maximum depending on how the algorithm is setup. While not the best solution, it may be good enough for the cost target. If a small output voltage range is considered, such as the one to a single battery which varies from 10 V to 15 V, the variation of the Windlar may be fine but the designer must decide if cost or precision is more important. The designer must also consider the speed at which the answer is required. The Wilson current mirror while accurate had a delay from 500 ns to 2 ms depending on bias currents where the Windlar had less than 200 ns. VREF = 3.3V R1 = 130k Current limit 100mV = 1A 200mV = 2A 300mV = 3A 400mV = 4A 500mV = 5A 600mV = 6A 700mV = 7A 800mV = 8A 900mV = 9A 1V = 10A SW1 R2 = 10.7k R3 = 18.7k SW2 R4 = 28k SW3 Figure 30. Output Charge Rate Programming Input and Output Current Balancing When building an ideal solar controller, the controller should protect the batteries or load while extracting the maximum energy from the solar panels. Unfortunately in the real world a customer or installer may purchase a large solar panel and a small battery. If the solar controller were to charge at the peak power, the battery would charge too fast and the lifetime of the battery would degrade or an explosion could occur. What the controller should do is to manage the needs of the battery and balance them with the peak power supplied from the solar panel. Thus a maximum battery charge rate programming as well as a voting scheme is needed to determine what is limiting the current out of the system. The programming of the current is accomplished via a 3.3 V reference provided by the NCP1294 and a resistor divider network. The shorting of one or multiple headers will result in a different current limit shown below where 100 mV is 1 A. 1 Short 0 s2 s3 Programmed Current 1 1 1 0.00 0 1 1 2.51 1 0 1 4.15 1 1 0 5.85 0 0 1 6.09 0 1 0 7.57 1 0 0 8.72 0 0 0 10.11 www.onsemi.com 22 Open s1 AND8490/D The user programmed set current is compared to the measured output current of the system and an error signal is produced. The error signal results in the voltage level set of the pulse by pulse current limiter. Likewise, the MPPT finds the maximum power point and sets the pulse by pulse current limiter for achieving the maximum power point. The lowest current set point should govern the system and the two signals will fight for dominance. If the two designs are merged into series regulators they can coexist. If a regulator is not able to achieve its intended higher current, it will keep opening the pass MOSFET until it is fully on and the other MOSFET will regulate the current. VREF = 3.3V VREF = 3.3V Output Current Amplifier Output Current Amplifier MPPT Current Limit + + Battery Current Limit − − Pulse by Pulse Current Limit Set Output Current Pulse by Pulse Current Limit Set Output Current VREF = 3.3V Output Current Amplifier + Battery Current Limit − Output Current Amplifier MPPT Current Limit Output Current + − Pulse by Pulse Current Limit Set Output Current Figure 31. Voting Scheme Solar Panel Connection Transients The operation of the circuit is shown graphically below where the maximum charge rate of the battery is set below the MPPT of the solar panel and the system tradeoff between charging at the maximum battery charge rate and the maximum power point is achieved. When connecting a solar panel to a solar controller, load voltage transient can occur and a designer must take action to protect the power supply from large transients. The voltage transients shown below for the panel described are about 14% above Voc. Current Maximum Power Point Current Limit Set Battery Maximum Charging Rate Output Current Time Figure 32. Battery Charging Between MPPT and Maximum Charge Rate www.onsemi.com 23 AND8490/D Figure 33. Crystal Solar Panel Connection to a 10 kW Resistive Load Positive Terminal Connected Last Voltage transients can be minimized in some of the following ways: suppression, filtering, or tolerating. If the designer chooses to suppress the transients, the power dissipated to suppress the transient must be calculated. The inductance of the connection wires can be calculated for a pair of wires as long d > 5*a using the following equation: L+ mO * l 4p ǒ H 3.95 mH + 4p * 10 −7 m * 9m a d l L mo 4p = = = = = ǒd *a aǓǓ * 1 ) 4 * ln ǒ * 1 ) 4 * ln ǒ (eq. 41) ǓǓ 7.5 mm * 1.5 mm 1.5 mm Wire radius Wire (center) distance Length of wire Wire inductance 4px10−7 Figure 34. Short Circuit Response of the Wurth Solar panel The proper part can be selected by looking at the pulse rating curve for the voltage suppressor. An example pulse rating curve is shown in Figure 35. The energy delivered to the load from the wire inductance is 1/2LI^2. The current delivered to the load will be at maximum short circuit current or Isc. From the scope capture shown in Figure 34, the transients do not last longer than 1 ms, thus the power dissipation needed for the transient voltage suppressor can be calculated as shown below: E+ 1 2 E Isc L TRTime * L * Isc 2 TR Time ³ 34.8 W + = = = = 1 * 3.95 2 2 mH * ǒ4 AǓ (eq. 42) 1 ms Energy Short circuit current Wire inductance Transient rise time www.onsemi.com 24 AND8490/D = = = = LFilter R1 R2 VSpike S Input inductance DCR of the input inductor Load resistance Resulting line spike LFilter V1 R1 Figure 36. Simplified Input Diagram If a reasonable filter is chosen where the inductance is 1 mH, capacitance is 10 mF, R1 is 1 mW and R2 is 1 MW the time base result of the filter can be calculated as shown in Figure 37. The amplitude of the voltage spike can be estimated by using v = l*di/dt. Assuming the inductance calculated above and the current is Isc, the remaining value to be obtained is the dt. The rise time of the current will vary with the technology, irradiance, and temperature of the solar panel and the wire inductance, but the reaction time of the Wurth solar panel with nine meters of wire connected is shown in Figure 34 and is less than 2 ms. 3.95 mH * 4.2 A * 0 A L * DI ³ 8.3 V + Dt 2 ms V Spike + DI L Dt VSpike = = = = 10 1 V2( t ) 0.1 (eq. 43) Change in current Wire inductance Time Resulting line spike 0.01 3 1.10 0 Since the transients seen on the input are short in duration, a filter can be used to reduce the amplitude of the voltage that the rest of the system will see. The worst case scenario is that the input system is charged up to the open circuit voltage of the solar panel with no load applied when the transient pulses are applied. According to [6] the voltage the system will be subjected to can be approximated by using the diagram shown in Figure 36 and equations below. a+ 1 R1 1 * ) 2 L Filter R2 * C Filter b+ Ǹ R1 ) R2 R2 * L Filter * C Filter V2(t) + a CFilter (eq. 44) 1* b * ǸL Filter * C Filter * e −a*t * sinǒb * t ) qǓ 6 1.5.10 2.10 6 Table 4. INPUT TO OUTPUT CONNECTION POSSIBILITIES V Spike * R2 ǒ 6 Aside from normal solar panel transients there also exists the possibility of four different input to output connections. b ǸR1)R2 R2 1.10 t Reverse Polarity Protection Case R1 ) R2 7 5.10 Figure 37. Change in Voltage Over Time at V2 * a2 ǒaǓ q + a tan R2 CFilter Vspike Figure 35. Pulse Rating Curve for the SMF5.0AT1 Series [5] V2 Output Voltage Input Voltage 1 Correct Correct 2 Correct Reverse 3 Reverse Correct 4 Reverse Reverse In case 1 there is no need for protection as the input and output are connected correctly. In case 2, the input voltage is connected in reverse. If current is allowed to flow in this case, then all but the output diode will have the possibility of being damaged. However, by placing a diode in series with the input either as shown in B or C in Figure 38 all devices can be protected. One drawback to the series diode is that it dissipates power continuously in a system. If the Ǔ = Exponential time constant = Input capacitance www.onsemi.com 25 AND8490/D diode is placed in for reverse polarity protection in a high current system, the losses can be significant. An alternative way to implement reverse polarity protection is to place a diode such that it opens a fuse when a reverse voltage is applied as shown in D in Figure 38. The fuse chosen can be a user replaceable or a polley thermal fuse. The fuse provides the necessary protection but can lead to an unfavorable user experience. A low loss way to implement the diode reverse polarity protection is to use a MOSFET that turns on when the voltage is applied in the proper polarity and turns off when it is not. One example of the appropriate circuit is shown in E in Figure 38. Damaged Damaged A Damaged Damaged Drive Protected Protected B Protected Drive A Protected Protected Protected C Protected Drive Protected Damaged Protected Protected D Protected Drive Protected Protected Protected E Protected Drive Protected Figure 38. Input Connected in Reverse Polarity components could be ignited from the large amount of energy. Figure 39 B shows one way to protect from reverse output voltage. In case 3 where the output is connected in reverse polarity and the input is connected correctly, three of the power components could be damaged. Since the source is assumed to be a lead acid battery, the protection is critical as damaged www.onsemi.com 26 AND8490/D Damaged A Damaged Drive Damaged Protected Protected B Protected Drive Figure 39. Output Connected in Reverse Polarity The final case is one in which both the input and output are connected incorrectly. In case 4, if the designer implemented both protections from case 2 and case 3 the input and output would be protected. The designer should not overlook the voltage suppressors that are installed on the input at the transient voltages which can be presented in either the proper polarity or reverse polarity. Therefore it is important to have bi−directional transient suppressors that can withstand the normal reverse polarity voltage without damage. Protected Protected Protected Protected Protected Protected Drive Figure 40. Input and Output Connected in Reverse Polarity One may wonder if the central grounding will affect the protection in place for reverse polarity protection and the answer is no. The topology implemented for protection in the design opens the path even with central grounding. Although the above methods serve as electrical means of preventing damage to the power supply, a system designer should not overlook a mechanical means of preventing damage such as a keyed connector or dissimilar connections. www.onsemi.com 27 AND8490/D Battery Drive Solar Panel Current Sense Resistor 10mW Central Grounding Point 1mW Battery Drive Solar Panel Current Sense Resistor 10mW Central Grounding Point 1mW Figure 41. Reverse Polarity Protection with Central Grounding Scheme Drive Scheme for Main Switches appropriate drive scheme to ensure good switching waveforms. The basic four switch buck boost converter is shown in Figure 42 but, Q1 and Q2 must be driven with the L1 D2 Q1 D1 Q2 Solar Panel Drive Battery Figure 42. General Schematic for Buck Boost change from a non−synchronous discrete driver to an integrated synchronous driver there are very few PMOSFET synchronous integrated drivers. If the designer manages to find a PMOSFET that is at cost parity with an NMOSFET and has the same gate charge, the chances are not good that there will be many second sources. Therefore, the choice of a PMOSFET may put the solar controller in a line down situation after production has started. For all of the aforementioned reasons, the high side MOSFET will be an NMOSFET but both a PMOSFET drive scheme and NMOSFET drive scheme are shown in the Figures 43 and 44 below. In the NMOSFET drive scheme shown in Figure 43, a drive signal is sent to the gate of Q1, in this case a 2N7002. When the drive signal from the NCP1294 is high, it pulls down on the R1 R2 resistor divider. The resistor divider in turn pulls down on the gate of Q3, turning it on and allowing C1 pre−charged 8 V to flow through the switching The next decision one would have to make is should Q1 be a PMOSFET or an NMOSFET. When making the Q1 design decision one must consider input voltage range, efficiency, price, and second sources. First, the input voltage range is 10 V to 60 V so the MOSFET must be able to handle a minimum of 60 V but keeping with good design practice, the break down voltage will need to be derated. In order to have a product with a long lifetime a part that is rated for 100V needs to be found. The converter should have the best possible efficiency as solar panels are currently expensive and the energy derived from them is precious. The MOSFET that should be selected has a low Rdson and low gate charge as the converter should switch as fast as possible to reduce the size of L1 which can be the largest part of a controller design. The price of a PMOSFET can be 2 to 4 times higher and the gate charge is usually double that of an NMOSFET with the same Rdson. Further, if the designer decides to www.onsemi.com 28 AND8490/D pulling down on VSW during Q2s off time to ensure C1 is properly charged. D3 and R5 can be used to change the nonoverlap time of the synchronous C1 charging. In looking at the drive scheme for the PMOSFETs, only 3 transistors are required versus the 7 needed for the N drive, but each transistor is approximately 0.03 USD in medium quantities, so quantitatively the P drive is 0.09 USD vs 0.21 USD for the N drive. The question one must ask is: is the PMOSFET for the application less than the cost difference of the drivers and will the efficiency be degraded? Another question is can a synchronous driver be used for the same cost as it will also improve efficiency? The answer is no because a driver that will work at 60 V will cost 0.71 USD and trading a diode for a MOSFET would double the cost for the switch. The main design will not be a synchronous design, but if the designer holds efficiency above cost, a driver has been made available so that the designer can prototype both circuits to weigh cost and efficiency. diode, D2, to the gate of Q2, turning it on and allowing the switch node to rise. As the switch node rises, so does the voltage at the gate of Q2 until the solar panel voltage is presented to the switch node and the solar panel voltage plus 8 V is presented at the gate of Q2. Once the gate signal goes low, the current ceases to flow through R1 and R2 and the gate of Q3 is pulled to VSW plus the bias voltage. R3 discharges to the switch node potential and Q4 discharges the gate of Q2 to the switch node. Upon completion of Q2 discharge the switch node falls and C1 charges up to 9 V through D1. In this type of drive scenario, there may be a startup into pre−bias issue where the converter never gets started. If the battery is connected first to the output, then the solar panel is connected, there will be 12 V at the switch node and 8 V on the VC node which will not allow capacitor C1 to charge properly. Since C1 is not charged properly, Q2 can only turn on linearly and will not start switching. The purpose of the pink highlighted area is to solve this issue by C1 = 0.1uF Q1 Gate NCP6412AN Q2 Solar Panel Voltage = 10V−60V MMSD4148 VC = 8.0V VSW D1 Q1 Drain R1 = 2k S G Q3 BSS84L Q3 Gate D R1 = 4k MMSD4148 R4 = 100 0V D2 2N7002 Q1 NCP1294 Gate D R3 = 499 Q4 Q3 Source MMBT3906 G 0V S 2N7002 MMSD4148 D3 Q5 Q2 Gate R5 = 499 Switch Node NMOS Drive Figure 43. NMOSFET Drive Scheme www.onsemi.com 29 AND8490/D NCP6412AN Q2 Solar Panel Voltage = 10V−60V VSW Q1 Gate Q3 MMBT3904 Q1 Drain R1 = 400 D1 MMSD4148 Q3 base 0V R2 = 200 2N7002 Q1 Q2 Gate D 0V G NCP1294 Gate S Switch Node Figure 44. PMOSFET Drive Scheme the full four switch buck boost will be engaged. The circuit in Figure 45 looks at the solar panel voltage and the battery charge voltage and determines when they have hit the predetermined duty cycle threshold. Once they have passed the 85% duty cycle, the full buck boost state will begin to operate. Q2 is only turned on during buck boost operation unlike Q1 which is driven continuously during both buck and buck boost operation. The solar controller should turn on the buck boost portion (Q2) of the circuit only when the input voltage falls too far to maintain regulation and that is determined by the choice of components and allowable max duty cycle. Therefore, a circuit is needed to determine at what duty cycle Duty Cycle Check Comparator Solar Panel R2 = 1.27k + LSDRV R1 = 10k − R3 = 1.0k Battery Voltage R4 = 10k L1 D2 Q1 R6 = 10 Battery Voltage D1 Q2 HSDRV R5 = 10k NCP1294 Gate R5 = 10k LSDRV www.onsemi.com 30 Battery AND8490/D Soalr panel Voltage Battery Votlage LSDRV Q4 Gate Q2 Gate Figure 45. Buck Boost Mode Detector Buck Mode Power Saving Option P MOSFET + R DS(on) ³ P MOSFET + I OUT 2R DS(on) ³ I OUT 2 (eq. 46) 7W 70 mW + 10 A 2 During buck mode, Q2 remains off and D2 is always forward conducting. In buck mode it is advantageous to short D2 since it does not serve a constructive purpose in this mode and the tradeoff made is one of efficiency for cost. The power dissipation of D2 is contained in the following equations. P diode + Vf * I OUT ³ 7 W + 0.7 * 10 A IOUT Pdiode Vf IOUT PMOSFET RDS(on) (eq. 45) = Output current = Power dissipation of a diode = Forward voltage of a diode = Output current = Power dissipation of the PMOSFET = On resistance of the PMOSFET The trade off is a good one as long as the Rdson is lower than 70 mW as shown graphically in the Figure 46 below, where the diode has a constant efficiency drop and the MOSFET losses have a linear loss factor. Figure 46. Efficiency Loss of Diode vs. MOSFET Battery Charging NCP1294 pulse by pulse current limiting and the current set circuit. The current will maintain the set charge rate configured by the designer or user unless the maximum power point falls below that level, at which time it will charge to the max power point adjusted rate. There are three stages to charging a lead acid battery: constant current or bulk charge, absorption or constant voltage mode, and float. During the bulk charge, the current is to be held constant which is accomplished with the www.onsemi.com 31 AND8490/D Figure 47. Lead Acid Battery Charge Profile the current drops to 5% of the amp hour rating. When the current drops to 5%, the solar controller will change to float state and the outer voltage loop will decrease to 13.5 V where it remains indefinitely or until the current need changes. Many types of lead acid batteries are suitable for solar powered street lights, they are: flooded, Valve Regulated Lead Acid (VRLA), Absorbed Gas Mat (AGM), and gel. A table of appropriate absorption, float and over voltage levels are shown in Table 5. Since the NCP1294 has pulse by pulse current limiting, nothing needs to change between Stage 1 and Stage 2 as the output voltage can be set to 14.5 V for both stages. In Stage 1, the outer voltage loop is never satisfied and the inner current loop maintains the required current. In Stage 2, the outer voltage loop takes over functionality and the inner current loop is not used except to maintain the maximum power point. The result of the voltage loop taking over is that a constant voltage of 14.5 V is maintained, while Table 5. CHARGING VOLTAGES OF LEAD ACID BATTERIES Voltage Range (V) Flooded Charging Mode Sealed VRLA AGM GEL Min Max Min Max Min Max Min Max Absorption 14.2 14.8 14.2 14.5 14.4 15 14.4 14.7 Float 13.2 13.5 13.2 13.5 13.2 13.8 13.5 13.8 Over Voltage 15.1 14.9 15.3 While the output voltage is set to 14.5 V for absorption and 13.5 V for float charge, the designer can place a potentiometer to dial in the voltage they would like for their particular battery. The voltage change circuitry is shown in Figure 48 below. 15.4 Battery Voltage R1 = 100 k OR NTC FB Voltage Output Current R2 = 80k or NTC R4 = 10k or NTC R3 = 69.8k R5 =330 or NTC Float Detection Comparator + − Float Current Trip Figure 48. Battery State Control and Internal Temperature Compensation www.onsemi.com 32 AND8490/D the FB node to be 1.263 V, therefore the resistance of R1 must be decreased or a temperature compensated current should be injected into the FB node. To determine how the NTC value changes over temperature, the following equation is used. One important thing to know is the affect temperature will have on battery charge voltage as most solar installations are not in sheltered environmental conditions. Typical lead acid battery voltage moves as an exponential. Some variation exists in the upper and lower battery tolerance at a specific temperature, the equations to define upper and lower battery voltages for the standard six cells configuration is shown below. Bat UPPER + 15.928617e −0.002479T Bat LOWER + 15.327247e −0.002582T BatUPPER BatLOWER T R + Ro b R Ro To T (eq. 47) (eq. 48) = Lower battery tolerance = Upper battery tolerance = Battery Temperature in degrees C = = = = = e b ǒT1*To1 Ǔ (eq. 49) Thermistor Constant Resistance at specified temperature Resistance at To 298.15 K (25°C) Current temperature in K If a NTC is placed in series with R1, the resulting battery voltage falls within the upper and lower battery limits established in Equations 47 and 48, except when it gets extremely cold. However, if a NTC replaces R2, the resulting curve does no harm in cold temperatures, but would destroy the battery when hot. When the two curves are put together, the resulting curve fits between the rails as shown below in Figures 49 and 50. To compensate for the battery change with temperature, the output set point of the solar controller must change with temperature. If remote sensing is not used, one idea for compensation is to use a negative temperature coefficient resistor or NTC as the voltage of the battery decreases as temperature increases. The NCP1294 regulates voltage at Figure 49. Temperature Compensated Battery Charged Voltage www.onsemi.com 33 AND8490/D Figure 50. Temperature Compensated Battery Float Voltage OOV The NCT loop should be placed as far away from the converter heat sources as possible. If the designer does not want to use an internal NTC, an external diode sense can be used in conjunction with the NCP1294 3.3 V reference. When using diodes to monitor the temperature as shown in Figures 49 and 51, it is important to buffer the 3.3 V reference voltage as the diode configuration can be remotely located on the battery or outside the case, where it will be exposed to ambient temperatures as well as ambient noise. Buffering the reference will isolate other circuitry which may be using the reference from the external noise contamination. The output battery voltage should be monitored to determine if the feedback mechanisms have been damaged or the remote sensing has influenced the battery voltage beyond the battery temperature compensation. The NCP1294 is equipped with a OOV comparator, when tripped the OOV shuts off the system. The comparator can be used on the input of the system or the output of the system, but is recommended to be used on the output as a failsafe mechanism. When using a single battery system, a single trip point of 18 V can be used or trip points can be set based on the state of charge. If using the float voltage state, a trip voltage of 15 V needs to be set. If using the absorption phase of charge, an 18 V level should be set. If a completely safe system were to be setup, a trip level could be set to 10% greater than the battery voltage adjusted for temperature and state of charge. External Remote Sensing Solar Controller OUV The converter input voltage should be monitored to determine if the input voltage level will cause a thermal issue. The NCP1294 is equipped with an under voltage lockout independent of VIN monitoring to ensure the input voltage is at the desired level for providing full output power. + 3.3V _ Battery Votlage OTP R1 = 225k NCP1294 Since the solar controller can be used in an inappropriate way, it is suggested that the temperature of the buck main switch be monitored to determine if it has exceed maximum temperature levels. If the temperature of the main MOSFET has exceeded appropriate levels, the current can be throttled back to reduce the power dissipation of the system. FB RBIAS = 25k R2 = 10k Figure 51. External Diode Temperature Monitoring www.onsemi.com 34 AND8490/D Thermal Management proper laboratory testing should be performed to ensure the design will dissipate the required power under worst case operating conditions. Variables considered during testing should include maximum ambient temperature, minimum airflow, maximum input voltage, maximum loading, and component variations (i.e., worst case MOSFET RDSON). Several layout tips are listed below for the best electric and thermal performance. The NCP1294 is a small source of power dissipation in the system for which the equations above detailed the loss mechanisms. The control portion of the IC power dissipation is determined by the formula below: P C + I CC ICC PC VIN V IN (eq. 50) = Control circuitry current draw = Control power dissipation = Input voltage The Solar Panel The solar panel supported by the NCP1294 evaluation board is between 5 W and 120 W. The industry standard types of solar panels were considered for this development. Most common types of solar cells are crystalline silicon in which there are two primary types: monocrystaline and polycrystalline silicon. Monacrystalline is the most efficient, but also more expensive to produce and is usually limited to commercial and residential applications. Amorphous solar panels consist of a thin film made from molten silicon that is spread across a large plate of stainless steel or similar material. The crystalline structures are very fragile and usually sandwiched between two sheets of glass for protection. • Monocrystalline 18% efficiency • Polycrystalline 15% efficiency • Amorphous 10% efficiency Once the IC power dissipations are determined, the designer can calculate the required thermal impedance to maintain a specified junction temperature at the worst case ambient temperature. The formula for calculating the junction temperature with the package in free air is: TJ + TA ) PD PD RqJA TA TJ R qJA (eq. 51) = Power dissipation of the IC = Thermal resistance junction to ambient of the regulator package = Ambient temperature = Junction temperature Thermal performance of the solar controller is strongly affected by the PCB layout. Extra care should be taken by users during the design process to ensure that the IC and power switches will operate under the recommended environmental conditions. As with any power design, Figure 52. Solar Controller PCB www.onsemi.com 35 36 Z3 Q22 NTD6415ANT4G HI www.onsemi.com Figure 53. Schematic 1Z GND2-1 Q6 MMBT3904LT1 ON 1k R57 R48 100k 14V 1k R56 D8 MMSD4148T3G 12 13 C39 10k R26 R5 78.7K 1.5nF C3 R20 100k 1uF IN AUX 10k R25 0.1uF C2 VC 1uF 560pF C1 ISETFB R27 13k 3P3 HI C8 1 OUT 1uF GND MC78L08ACP 2 U11 C4 205 kHz 95.25% R19 100k C16 3 ON 680uF 63V 680uF 63V 8 7 6 5 HSSENS LM5109A U4 VDD HB HO HI HS LI VSS LO R6 1 2 3 4 10m C15 VC C14 UVLO set to 7.86V OOV set to 17.74V 0.1uF R46 4.99k R47 100k HI 2N7002KT1G Q20 14V 12V-52V 14 12V-52V U2-D + LM324AM 10k R53 1 2 3 4 5 6 7 8 0.1uF Q7 C12 VC PGND VCC VREF LGND SS COMP VFB U1 NCP1294 GATE ISENSE SYNC FF UV OV RTCT ISET 10R R15 VCC 16 15 14 13 12 11 10 9 VC R17 4k R13 2k LO 2N7002KT1G HB HO VSW1 10k R60 3P3BUFF 330nF C23 VCC 3P3 200k 1nF C6 MMSD4148T3G D7 R41 D2 14V 2N7002KT1G Q8 Q1 IREG 1R0 10k R9 2N7002KT1G Q12 R8 BIPASS LO LSDRV 1R0 R59 MMBT3906LT1 MMSD4148T3G R14 10k R42 100R0 R18 499R Q18 HO Q3 3 BSS84LT1 D S 2 Q14 AOT414 -20%, +80% 25V C24 0.1uF 2N7002KT1G 1G D5 MMSD4148T3G HB VC VSW1 22uH L1 C11 C7 220nF AOT414 Q13 150pF 100k R7 R11 100R0 1nF 3 1 2 R32 680uF 63V C9 10m GND1 VFM2 VFM1 R3 20R FLOAT NT3 100k NTC Output Current Sense 2 ATP104 Q21 MBR30H100CT D6 3 1 BIPASS C13 SER2918H-223KL D1 MBR30H100CT 2 10k R58 G1 S 2 Q10 BSS84LT1 3 D 1 10k R55 R54 10k 3 VSW1 C10 2N7002KT1G Q9 R4 69.8k NT2 80k NTC R2 499R 1.5nF R1 330 NTC NT1 10k NTC 14V 12V-52V SOALR- Q19 NTD6415ANT4G 14V SOLAR+ - BAT+ BAT AND8490/D Z1 AND8490/D MPP-1 12V-52V 12V-52V MPP-2 VC MPP-3 VC IREG MPP-4 IREG 3P3 3P3 MPP-5 MPP-6 MPPFB MPP-7 VC 10k 6 R52 5 R50 U2-B + 7 LSDRV LM324AM 10k R51 1.27k - 12V-52V 14V R49 1k Figure 54. Schematic 2 www.onsemi.com 37 C20 C19 C18 C17 C26 0.1uF 0.1uF 0.1uF 0.1uF 0.1uF AND8490/D 12V-52V 14V VOUT_FILT 3P3 R39 49.9 HSSENS R40 499R 3P3BUFF BC846BPDW1T1G Q11 Q4 Q5 Q15 5 5 2 2 3 IMPPDRV 1 IREG 1 BC846BPDW1T1G 1 2 1 2 IBATDRV 6 6 6 6 Q17 2N7002KT1G BC846BPDW1T1G 5 3 5 3 IMPPDRV 2N7002KT1G BC846BPDW1T1G 3 1 U2-A + LM324AM 4 - 4 3 4 3P3 4 2 Q16 IOUT ISETFB C22 3.9nF R43 4.99k R21 1k LM324AM 8 IMPPDRV 10 MPPFB + U2-C 9 - R38 4.99k R35 3P3BUFF R10 3.9nF 499R MPPFB 14V C5 R33 499R 1k R29 R16 10k R45 100k 10k R31 Q2 3 10k 1uF C42 D4 R30 MMSD4148T3G ON IBATDRV 100k R37 2 1 LM324AM 1 0R0 ISETFB U3-A 3 + 10k R24 5 + U3-B 6 - 1uF BSTEN GND FLOAT LM324AM 8 IO3-1 IOUT C25 0.1uF R12 C21 2N7002KT1G 2 2.8k R44 LM324AM 7 D3 0R0 IO3-2 10 + U3-C 9 - 2% Trip R36 10.7k R28 100k 14 U3-D + LM324AM R34 28k 13 12 R23 18.7k IO2-1 IO2-2 IO1-1 IO1-2 R22 28k Figure 55. Schematic 3 www.onsemi.com 38 IO1 IO2 IO3 0 0 0.91 11.4 0 0 0 1 V I .78 9.7 0 1 0 .67 8.1 0 1 1 .51 6 1 0 0 .55 6.5 1 0 1 0.37 4.1 1 1 1 1 0 1 .22 0 2.0 0 AND8490/D Figure 56. Layout Bottom www.onsemi.com 39 AND8490/D Figure 57. Layout Top Table 6. BOM Reference Qty Description Part Name Manufacturer Footprint Manufacturer Manufacturer Part # Q2, Q7 − 9, Q12, Q16 − 18, Q20 9 Small Signal N MOSFET 60 V 380 mA NA SOT−23 ON Semiconductor 2N7002KT1G Q21 1 30 V, 72 A, Single P−Channel MOSFET 8.4 mW NA ATPAK (2leads + tab) SANYO ATP104−TL−H D2, D4 − 5, D7−8 5 Switching Diodes 1V NA SOD123 ON Semiconductor MMSD4148T3G Q4 − 5, Q11, Q15 4 NPN Dual 65 V 100 mA NA SC−88 ON Semiconductor BC846BPDW1T1G Q3, Q10 2 NFET 10R 50 V NA SOT−23 ON Semiconductor BSS84LT1 C1 1 SMT Ceramic Capacitor 560 pF ±5% 603 TDK Corporation C1608C0G1H561J C3, C10 2 SMT Ceramic Capacitor 1.5 nF ±10% 603 TDK Corporation C1608X7R2A152K www.onsemi.com 40 AND8490/D Table 6. BOM Part Name Manufacturer Footprint Manufacturer Manufacturer Part # SMT Ceramic Capacitor 1 nF 10% 603 TDK C1608X7R2A102K 2 SMT Ceramic Capacitor 1 mF −20%, +80% 603 Taiyo Yuden EMK107B7105KA−T C23 1 SMT Ceramic Capacitor 330 nF ±10% 603 TDK Corporation C1608X7R1A334K C17−20, C25−26 6 SMT Capacitor 0.1 mF ±20% 805 AVX Corporation 08055C104MAT2A C2, C12, C24 3 Ceramic Chip Capacitor 0.1 mF −20%, +80% 603 AVX 06033G104ZAT2A C11 1 Ceramic Chip Capacitor 220 nF 10% 603 TDK Corporation C1608X7R1C224K C5, C22 2 Ceramic Chip Capacitor 3.9 nF 10% 603 Murata GRM188R71H392KA01D C21, C42 2 Ceramic Chip Capacitor 1 mF 10% 603 Murata GRM188R71H392KA01D C39 1 Ceramic Chip Capacitor 0.1 mF ±20% 603 TDK Corporation C1608X7S2A104M C7 1 Ceramic Chip Capacitor 150 pF 5% 603 TDK Corporation C1608C0G1H151J C16 1 Ceramic Chip Capacitor 1 mF ±10% 1206 Murata GRM31MF51E105ZA01L U1 1 Enhanced Voltage Mode PWM Controller 3V Reference NA SOIC 16 ON Semiconductor NCP1294 C9, C14−15 3 Electrolytic Capacitor 680 mF 63 V 20% 16X25 Nichicon UPW1J471MHD6 U2−3 2 QUAD, LOW POWER OP AMP 1 MHz NA 14−SOIC ON Semiconductor LM324ADR2G U4 1 Half Bridge Driver 1 A 108 V NA 8−WDFN National Semiconductor LM5109AMA D1, D6 2 Schottky Rectifier 100 V 30 A NA TO−220 ON Semiconductor MBR30H100CTG U11 1 8.0 V Regulator 4% TO−92−3 ON Semiconductor MC78L08ACPRE Q6 1 General Purpose NPN Transistor 40 V 200 mA NA SOT−23 ON Semiconductor MMBT3904LT1G Q1 1 General Purpose PNP Transistor PNP NA SOT−23 ON Semiconductor MMBT3906LT1G Reference Qty Description C6, C13 2 C4, C8 Z1, Z3 2 Q13−14 2 N−MOSFET 100 V 18.2m NA TO−220 Alpha & Omega AOT414 NI Q19, Q22 2 N MOSFET 100 V 50m NA Dpak ON Semiconductor NTD6415ANT4G NT1 1 Resistor 10k NTC ±1.0% 603 EPCOS Inc B57331V2103J60 NT2 1 Resistor 80k NTC ±1.0% 603 Vishay / Dale NTHS0603N01N8002JE NT3 1 Resistor 100k NTC ±1.0% 603 TDK NTCG164KF104FT1 R1 1 Resistor 330 NTC ±1.0% 603 Murata NCP18XM331J03RB R2, R18, R33, R35, R40 5 Resistor 499R ±1.0% 603 Stackpole Electronics Inc RMCF0603FT499R R9, R14, R24−26, R30−31, R45, R50, R52 − 55, R58, R60 15 Resistor 10k ±1.0% 603 Panasonic ERJ−3EKF1002V RMCF0603FT2K00 R13 1 Resistor 2k ±1.0% 603 Stackpole Electronics Inc R10, R49, R21 3 Resistor 1k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT1K00 R29, R12 2 Resistor 0R0 ±1.0% 603 Stackpole Electronics Inc RMCF0603FT499R R16, R37, R7, R47 − 48, R19 − 20, R28 8 Resistor 100k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT100K R34, R22 2 Resistor 28k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT28K0 R43, R46, R38 3 Resistor 4.99k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT4K99 R44 1 Resistor 2.8k ±1.0% 603 Vishay / Dale RMCF0603FT2k80 R5 1 Resistor 78.7k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT78K7 www.onsemi.com 41 AND8490/D Table 6. BOM Reference Qty Description Part Name Manufacturer Footprint R36 1 Resistor 10.7k ±1.0% R27 1 Resistor 13k ±1.0% R3 1 Resistor 20R R51 1 Resistor R4 1 Resistor R17 1 R39 1 R41 Manufacturer Manufacturer Part # 603 STACKPOLE RMCF0603FT10K7 603 Stackpole Electronics Inc RMCF0603FT13K0 ±1.0% 603 Stackpole Electronics Inc RMCF0603FT20R0 1.27k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT1K27 69.8k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT69K8 Resistor 4k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT4K99 Resistor 49.9 ±1.0% 603 Stackpole Electronics Inc RMCF0603FT49R9 1 Resistor 200k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT4K99 R56−57 2 Resistor 1k ±1.0% 603 Panasonic ERJ−3EKF1002V R23 1 Resistor 18.7k ±1.0% 603 Stackpole Electronics Inc RMCF0603FT15K0 R15 1 Resistor 10R ±5.0% 603 Stackpole Electronics Inc RMCF0603JT10R0 RMCF1206JT1R00 R8, R59 2 Resistor 1R0 ±5.0% 1206 Stackpole Electronics Inc R42, R11 2 Resistor 100R0 ±5.0% 1206 Stackpole Electronics Inc RMCF1206JT100R R6, R32 2 Resistor 10m ±1.0% 2512 Bourns Inc. CRA2512−FZ−R010ELF L1 1 Inductor 22 mH ±10% SMT Coilcraft SER2918H−223KL D3 1 Shunt Reference ON Semiconductor TL431ACLPRAG ±1.0% Bibliography 1. BP 3220T Datasheet.” BP Solar USA. BP Solar, 01 Jan. 2010. Web. 21 Mar. 2011. <http://www.bp.com/liveassets/bp_internet/solar/bp_solar_usa/STAGING/local_assets/downloads_pdfs/pq/BP3220T _1−10.pdf>. 2. T. Esram and P. L. Chapman, “Comparison of photovoltaic array maximum power point tracking techniques,” IEEE Trans. Energy Conv., Vol. 22, pp. 439–449, 2007. 3. [“Irradiance − Definition and More from the Free Merriam−Webster Dictionary.” Dictionary and Thesaurus − Merriam−Webster Online. Web. 21 Mar. 2011. <http://www.merriam−webster.com/dictionary/irradiance>. 4. “NCP1294 Datasheet.” ON Semiconductor Home Page. ON Semiconductor, July 2090. Web. 21 Mar. 2011. <http://www.onsemi.com/pub_link/Collateral/NCP1294−D.PDF>. 5. “SMF5.0AT1 Series Datasheet.” ON Semiconductor Home Page. ON Semiconductor, Sept. 2007. Web. 21 Mar. 2011. <http://www.onsemi.com/pub_link/Collateral/SMF5.0AT1−D.PDF>. 6. Tarter, Ralph E. The RLCR Circuit with a Dc Input. Solid−state Power Conversion Handbook. New York: Wiley, 1993. 26−27. Print. ON Semiconductor and the are registered trademarks of Semiconductor Components Industries, LLC (SCILLC) or its subsidiaries in the United States and/or other countries. SCILLC owns the rights to a number of patents, trademarks, copyrights, trade secrets, and other intellectual property. A listing of SCILLC’s product/patent coverage may be accessed at www.onsemi.com/site/pdf/Patent−Marking.pdf. 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 19521 E. 32nd Pkwy, Aurora, Colorado 80011 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: 81−3−5817−1050 www.onsemi.com 42 ON Semiconductor Website: www.onsemi.com Order Literature: http://www.onsemi.com/orderlit For additional information, please contact your local Sales Representative AND8490/D