LM43603-Q1 3.5V 至 36V、3A 同步降压转换器

7.3.1 Fixed Frequency Peak Current Mode Controlled Step-Down Regulator

The following operating description of the LM43603-Q1 refers to the Functional Block Diagram and to the waveforms in Figure 35. The LM43603-Q1 is a step-down buck regulator with both a high-side (HS) and low-side (LS) switch integrated into the device. The LM43603-Q1 supplies a regulated output voltage by turning on the HS and LS NMOS switches with controlled ON time. During the HS switch ON time, the SW pin voltage V_SW swings up to approximately V_IN, and the inductor current i_L increases with a linear slope (V_IN – V_OUT) / L. When the HS switch is turned off by the control logic, the LS switch is turned on after a anti-shoot-through dead time. Inductor current discharges through the LS switch with a slope of –V_OUT / L. The control parameter of buck converters are defined as duty cycle D = t_ON / T_SW, where t_ON is the HS switch ON time and T_SW is the switching period. The regulator control loop maintains a constant output voltage by adjusting the duty cycle D. In an ideal buck converter, where losses are ignored, D is proportional to the output voltage and inversely proportional to the input voltage: D = V_OUT / V_IN.

Figure 35. SW Node and Inductor Current Waveforms in Continuous Conduction Mode (CCM)

The LM43603-Q1 synchronous buck converter employs peak current mode control topology. A voltage feedback loop is used to get accurate DC voltage regulation by adjusting the peak current command based on voltage offset. The peak inductor current is sensed from the HS switch and compared to the peak current to control the ON time of the HS switch. The voltage feedback loop is internally compensated, which allows for fewer external components, makes it easy to design, and provides stable operation with almost any combination of output capacitors. The regulator operates with fixed switching frequency in CCM and DCM. At very light load, the LM43603-Q1 operates in PFM to maintain high efficiency and the switching frequency decreases with reduced load current.

7.3.2 Light Load Operation

DCM operation is employed in the LM43603-Q1 when the inductor current valley reaches zero. The LM43603-Q1 is in DCM when load current is less than half of the peak-to-peak inductor current ripple in CCM. In DCM, the LS switch is turned off when the inductor current reaches zero. Switching loss is reduced by turning off the LS FET at zero current, and the conduction loss is lowered by not allowing negative current conduction. Power conversion efficiency is higher in DCM than CCM under the same conditions.

In DCM, the HS switch ON time reduces with lower load current. When either the minimum HS switch ON time (t_ON-MIN) or the minimum peak inductor current (I_PEAK-MIN) is reached, the switching frequency decreases to maintain regulation. At this point, the LM43603-Q1 operates in PFM. In PFM, switching frequency is decreased by the control loop when load current reduces to maintain output voltage regulation. Switching loss is further reduced in PFM operation due to less frequent switching actions.

In PFM operation, a small positive DC offset is required at the output voltage to activate the PFM detector. The lower the frequency in PFM, the more DC offset is needed at V_OUT. Refer to the Typical Characteristics for typical DC offset at very light load. If the DC offset on V_OUT is not acceptable for a given application, a static load at output is recommended to reduce or eliminate the offset. Lowering values of the feedback divider R_FBT and R_FBB can also serve as a static load. In conditions with low V_IN and/or high frequency, the LM43603-Q1 may not enter PFM mode if the output voltage cannot be charged up to provide the trigger to activate the PFM detector. Once the LM43603-Q1 is operating in PFM mode at higher V_IN, it remains in PFM operation when V_IN is reduced. See Figure 45 for a sample of PFM operation.

7.3.3 Adjustable Output Voltage

The voltage regulation loop in the LM43603-Q1 regulates output voltage by maintaining the voltage on FB pin (V_FB) to be the same as the internal REF voltage (V_REF). A resistor divider pair is needed to program the ratio from output voltage V_OUT to V_FB. The resistor divider is connected from the V_OUT of the LM43603-Q1 to ground with the mid-point connecting to the FB pin.

Figure 36. Output Voltage Setting

The voltage reference system produces a precise voltage reference over temperature. The internal REF voltage is 1.011 V typically. To program the output voltage of the LM43603-Q1 to be a certain value V_OUT, R_FBB can be calculated with a selected R_FBT by Equation 1:

Equation 1.

The choice of the R_FBT depends on the application. TI recommends R_FBT in the range from 10 kΩ to 100 kΩ for most applications. A lower R_FBT value can be used if static loading is desired to reduce V_OUT offset in PFM operation. Lower R_FBT reduces efficiency at very light load. Less static current goes through a larger R_FBT and might be more desirable when light load efficiency is critical. But R_FBT larger than 1 MΩ is not recommended because it makes the feedback path more susceptible to noise. Larger R_FBT value requires more carefully designed feedback path on the PCB. The tolerance and temperature variation of the resistor dividers affect the output voltage regulation. TI recommends using divider resistors with 1% tolerance or better and temperature coefficient of 100 ppm or lower.

If the resistor divider is not connected properly, output voltage cannot be regulated because the feedback loop is broken. If the FB pin is shorted to ground, the output voltage is driven close to V_IN, because the regulator sees very low voltage on the FB pin and tries to regulator it up. The load connected to the output could be damaged under such a condition. Do not short FB pin to ground when the LM43603-Q1 is enabled. It is important to route the feedback trace away from the noisy area of the PCB. For more layout recommendations, see the Layout section.

7.3.4 Enable (EN)

Voltage on the EN pin (V_EN) controls the ON or OFF operation of the LM43603-Q1. Applying a voltage less than 0.4 V to the EN input shuts down the operation of the LM43603-Q1. In shutdown mode the quiescent current drops to typically 1.2 µA at V_IN = 12 V.

The internal LDO output voltage V_CC is turned on when V_EN is higher than 1.2 V. Switching action and output regulation are enabled when V_EN is greater than 2.1 V (typical). The LM43603-Q1 supplies regulated output voltage when enabled and output current up to 3 A.

The EN pin is an input and cannot be open circuit or floating. The simplest way to enable the operation of the LM43603-Q1 is to connect the EN pin to VIN pins directly. This allows self-start-up when V_IN is within the operation range.

Many applications benefit from use of an enable divider R_ENT and R_ENB in Figure 37 to establish a precision system UVLO level for the stage. System UVLO can be used for supplies operating from utility power as well as battery power. It can be used for sequencing, ensuring reliable operation, or supply protection, such as a battery discharge level. An external logic signal can also be used to drive EN input for system sequencing and protection.

Figure 37. System UVLO by Enable Dividers

7.3.5 VCC, UVLO, and BIAS

The LM43603-Q1 integrates an internal LDO to generate V_CC for control circuitry and MOSFET drivers. The nominal voltage for V_CC is 3.28 V. The VCC pin is the output of the LDO must be properly bypassed. Place a high-quality ceramic capacitor with 2.2-µF to 10-µF capacitance and 6.3 V or higher rated voltage as possible to VCC and grounded to the exposed PAD and ground pins. The VCC output pin must not be loaded, left floating, or shorted to ground during operation. Shorting VCC to ground during operation may cause damage to the LM43603-Q1.

Undervoltage lockout (UVLO) prevents the LM43603-Q1 from operating until the V_CC voltage exceeds 3.1 V (typical). The V_CC UVLO threshold has 520 mV of hysteresis (typically) to prevent undesired shuting down due to temporary V_IN droops.

The internal LDO has two inputs: primary from VIN and secondary from BIAS input. The BIAS input powers the LDO when V_BIAS is higher than the change-over threshold. Power loss of an LDO is calculated by I_LDO × (V_IN-LDO – V_OUT-LDO). The higher the difference between the input and output voltages of the LDO, the more power loss occur to supply the same output current. The BIAS input is designed to reduce the difference of the input and output voltages of the LDO to reduce power loss and improve LM43603-Q1 efficiency, especially at light load. It is recommended to tie the BIAS pin to V_OUT when V_OUT ≥ 3.3 V. Ground the BIAS pin in applications with V_OUT less than 3.3 V. BIAS input can also come from an external voltage source, if available, to reduce power loss. When used, TI recommends a 1-µF to 10-µF high-quality ceramic capacitor to bypass the BIAS pin to ground.

7.3.6 Soft-Start and Voltage Tracking (SS/TRK)

The LM43603-Q1 has a flexible and easy-to-use start-up rate control pin: SS/TRK. Soft-start feature is to prevent inrush current impacting the LM43603-Q1 and its supply when power is first applied. Soft start is achieved by slowly ramping up the target regulation voltage when the device is first enabled or powered up.

The simplest way to use the part is to leave the SS/TRK pin open circuit or floating. The LM43603-Q1 employs the internal soft-start control ramp and starts up to the regulated output voltage in 4.1 ms typically.

In applications with a large amount of output capacitors, or higher V_OUT, or other special requirements the soft-start time can be extended by connecting an external capacitor C_SS from SS/TRK pin to AGND. Extended soft-start time further reduces the supply current needed to charge up output capacitors and supply any output loading. An internal current source (I_SSC = 2 µA) charges C_SS and generates a ramp from 0 V to V_FB to control the ramp-up rate of the output voltage. For a desired soft start time t_SS, the capacitance for C_SS can be found with Equation 2:

Equation 2.

The LM43603-Q1 is capable of starting up into prebiased output conditions. When the inductor current reaches zero, the LS switch is turned off to avoid negative current conduction. This operation mode is also called diode emulation mode. It is built-in by the DCM operation in light loads. With a prebiased output voltage, the LM43603-Q1 waits until the soft-start ramp allows regulation above the prebiased voltage and then follows the soft-start ramp to the regulation level.

When an external voltage ramp is applied to the SS/TRK pin, the LM43603-Q1 FB voltage follows the ramp if the ramp magnitude is lower than the internal soft-start ramp. A resistor divider pair can be used on the external control ramp to the SS/TRK pin to program the tracking rate of the output voltage. The final voltage seen by the SS/TRK pin should not fall below 1.2 V to avoid abnormal operation.

Figure 38. Soft Start Tracking External Ramp

V_OUT tracked to external voltage ramps has options of ramping up slower or faster than the internal voltage ramp. V_FB always follows the lower potential of the internal voltage ramp and the voltage on the SS/TRK pin. Figure 39 shows the case when V_OUT ramps slower than the internal ramp, while Figure 40 shows when V_OUT ramps faster than the internal ramp. Faster start-up time may result in inductor current tripping current protection during start-up. Use with special care.

Figure 39. Tracking with Longer Start-up Time than the Internal Ramp

Figure 40. Tracking with Shorter Start-up Time than the Internal Ramp

7.3.7 Switching Frequency (RT) and Synchronization (SYNC)

The switching frequency of the LM43603-Q1 can be programmed by the impedance R_T from the RT pin to ground. The frequency is inversely proportional to the R_T resistance. The RT pin can be left floating, and the LM43603-Q1 operates at 500-kHz default switching frequency. The RT pin is not designed to be shorted to ground. For a desired frequency, typical R_T resistance can be found by Equation 3. Table 1 gives typical R_T values for a given F_S.

Figure 41. RT vs Frequency Curve

The LM43603-Q1 switching action can also be synchronized to an external clock from 200 kHz to 2.2 MHz. Connect an external clock to the SYNC pin, with proper high-speed termination, to avoid ringing. Ground the SYNC pin if not used.

Figure 42. Frequency Synchronization

The recommendations for the external clock include high level no lower than 2 V, low level no higher than 0.4 V, duty cycle between 10% and 90%, and both positive and negative pulse width no shorter than 80 ns. When the external clock fails at logic high or low, the LM43603-Q1 switches at the frequency programmed by the R_T resistor after a time-out period. TI recommends connecting a resistor R_T to the RT pin so that the internal oscillator frequency is the same as the target clock frequency when the LM43603-Q1 is synchronized to an external clock. This allows the regulator to continue operating at approximately the same switching frequency if the external clock fails.

The choice of switching frequency is usually a compromise between conversion efficiency and the size of the circuit. Lower switching frequency implies reduced switching losses (including gate charge losses, switch transition losses, etc.) and usually results in higher overall efficiency. However, higher switching frequency allows use of smaller LC output filters and hence a more compact design. Lower inductance also helps transient response (higher large signal slew rate of inductor current), and reduces the DCR loss. The optimal switching frequency is usually a trade-off in a given application and thus needs to be determined on a case-by-case basis. It is related to the input voltage, output voltage, most frequent load current level(s), external component choices, and circuit size requirement. The choice of switching frequency may also be limited if an operating condition triggers T_ON-MIN or T_OFF-MIN.

7.3.8 Minimum ON Time, Minimum OFF Time and Frequency Foldback at Dropout Conditions

Minimum ON time, T_ON-MIN, is the smallest duration of time that the HS switch can be on. T_ON-MIN value is typically 125 ns in the LM43603-Q1. Minimum OFF time, T_OFF-MIN, is the smallest duration that the HS switch can be off. T_OFF-MIN value is typically 200 ns in the LM43603-Q1.

In CCM operation, T_ON-MIN and T_OFF-MIN limits the voltage conversion range given a selected switching frequency. The minimum duty cycle allowed is:

And the maximum duty cycle allowed is:

Given fixed T_ON-MIN and T_OFF-MIN, the higher the switching frequency the narrower the range of the allowed duty cycle. In the LM43603-Q1, frequency foldback scheme is employed to extend the maximum duty cycle when T_OFF-MIN is reached. The switching frequency decreases once longer duty cycle is needed under low V_IN conditions. The switching frequency can be decreased to approximately 1/10 of the programmed frequency by R_T or the synchronization clock. Such wide range of frequency foldback allows the LM43603-Q1 output voltage stay in regulation with a much lower supply voltage V_IN. This leads to a lower effective dropout voltage. See Typical Characteristics for more details.

Given an output voltage, the choice of the switching frequency affects the allowed input voltage range, solution size and efficiency. The maximum operatable supply voltage can be found by:

At lower supply voltage, the switching frequency decreases once T_OFF-MIN is tripped. The minimum V_IN without frequency foldback can be approximated by Equation 7:

Taking considerations of power losses in the system with heavy load operation, V_IN-MIN is higher than the result calculated in Equation 7. With frequency foldback, V_IN-MIN is lowered by decreased F_S.

Figure 43. V_OUT = 5 V Fs = 500 kHz
Frequency Foldback at Dropout

7.3.9 Internal Compensation and C_FF

The LM43603-Q1 is internally compensated with R_C = 400 kΩ and C_C = 50 pF as shown in Functional Block Diagram. The internal compensation is designed such that the loop response is stable over the entire operating frequency and output voltage range. Depending on the output voltage, the compensation loop phase margin can be low with all ceramic capacitors. TI recommends an external feed-forward capacitor, C_FF, be placed in parallel with the top resistor divider R_FBT for optimum transient performance.

Figure 44. Feed-Forward Capacitor for Loop Compensation

The feed-forward capacitor C_FF in parallel with R_FBT places an additional zero before the cross over frequency of the control loop to boost phase margin. The zero frequency can be found with Equation 8:

An additional pole is also introduced with C_FF at the frequency of

Select the C_FF so that the bandwidth of the control loop without the C_FF is centered between f_Z-CFF and f_P-CFF. The zero f_Z-CFF adds phase boost at the crossover frequency and improves transient response. The pole f_P-CFF helps maintaining proper gain margin at frequency beyond the crossover.

Designs with different combinations of output capacitors need different C_FF. Different types of capacitors have different equivalent series resistance (ESR). Ceramic capacitors have the smallest ESR and need the most C_FF. Electrolytic capacitors have much larger ESR

and the ESR zero frequency would be low enough to boost the phase up around the crossover frequency. Designs using mostly electrolytic capacitors at the output may not need any C_FF.

The C_FF creates a time constant with R_FBT that couples in the attenuated output voltage ripple to the FB node. If the C_FF value is too large, it can couple too much ripple to the FB and affect V_OUT regulation. It could also couple too much transient voltage deviation and falsely trip PGOOD thresholds. Therefore, calculate C_FF based on output capacitors used in the system. At cold temperatures, the value of C_FF might change based on the tolerance of the chosen component. This may reduce its impedance and ease noise coupling on the FB node. To avoid this, more capacitance can be added to the output or the value of C_FF can be reduced. See Detailed Design Procedure for the calculation of C_FF.

7.3.10 Bootstrap Voltage (BOOT)

The driver of the HS switch requires a bias voltage higher than V_IN when the HS switch is ON. The capacitor connected between CBOOT and SW pins works as a charge pump to boost voltage on the CBOOT pin to (V_SW + V_CC). The boot diode is integrated on the LM43603-Q1 die to minimize the bill of material (BOM). A synchronous switch is also integrated in parallel with the boot diode to reduce voltage drop on CBOOT. A high-quality ceramic 0.47 µF, 6.3 V or higher capacitor is recommended for C_BOOT.

7.3.11 Power Good (PGOOD)

The LM43603-Q1 has a built-in power-good flag shown on PGOOD pin to indicate whether the output voltage is within its regulation level. The PGOOD signal can be used for start-up sequencing of multiple rails or fault protection. The PGOOD pin is an open-drain output that requires a pullup resistor to an appropriate DC voltage. Voltage detected by the PGOOD pin must never exceed 12 V. A resistor divider pair can be used to divide the voltage down from a higher potential. A typical range of pullup resistor value is 10 kΩ to 100 kΩ.

When the FB voltage is within the power-good band, +4% above a –7% below the internal reference V_REF typically, the PGOOD switch will be turned off, and the PGOOD voltage will be pulled up to the voltage level defined by the pullup resistor or divider. When the FB voltage is outside of the tolerance band, +10% above or –13% below V_REF typically, the PGOOD switch turns on, and the PGOOD pin voltage will be pulled low to indicate power bad. Both rising and falling edges of the power-good flag have a built-in 220 µs (typical) deglitch delay.

7.3.12 Overcurrent and Short-Circuit Protection

The LM43603-Q1 is protected from overcurrent conditions by cycle-by-cycle current limiting on both the peak and valley of the inductor current. Hiccup mode is activated if a fault condition persists to prevent over heating.

High-side MOSFET overcurrent protection is implemented by the nature of the peak current mode control. The HS switch current is sensed when the HS is turned on after a set blanking time. The HS switch current is compared to the output of the error amplifier (EA) minus slope compensation every switching cycle. Refer to Functional Block Diagram for more details. The peak current of the HS switch is limited by the maximum EA output voltage minus the slope compensation at every switching cycle. The slope compensation magnitude at the peak current is proportional to the duty cycle.

When the LS switch is turned on, the current going through it is also sensed and monitored. The LS switch is not turned OFF at the end of a switching cycle if its current is above the LS current limit I_LS-LIMIT. The LS switch is kept ON so that inductor current keeps ramping down, until the inductor current ramps below the LS current limit. Then the LS switch is turned OFF, and the HS switch is turned on, after a dead time. If the current of the LS switch is higher than the LS current limit for 32 consecutive cycles, and the power-good flag is low, hiccup current protection mode is activated. In hiccup mode, the regulator is shut down and kept off for 5.5 ms typically before the LM43603-Q1 tries to start again. If an overcurrent or short-circuit fault condition still exists, hiccup repeats until the fault condition is removed. Hiccup mode reduces power dissipation under severe overcurrent conditions, prevents overheating, and potential damage to the device.

Hiccup is only activated when power-good flag is low. Under non-severe overcurrent conditions when V_OUT has not fallen outside of the PGOOD tolerance band, the LM43603-Q1 reduces the switching frequency and keep the inductor current valley clamped at the LS current limit level. This operation mode allows slight over current operation during load transients without tripping hiccup. If the power-good flag becomes low, hiccup operation starts after LS current limit is tripped 32 consecutive cycles.

7.3.13 Thermal Shutdown

Thermal shutdown is a built-in self protection to limit junction temperature and prevent damage due to overheating. Thermal shutdown turns off the device when the junction temperature exceeds 160°C typically to prevent further power dissipation and temperature rise. Junction temperature reduces after thermal shutdown. The LM43603-Q1 restarts when the junction temperature drops to 150°C.

7.4.1 Shutdown Mode

The EN pin provides electrical ON and OFF control for the LM43603-Q1. When V_EN is below 0.4 V, the device is in shutdown mode. Both the internal LDO and the switching regulator are off. In shutdown mode the quiescent current drops to 1.2 µA typically with V_IN = 12 V. The LM43603-Q1 also employs undervoltage lockout protection. If V_CC voltage is below the UVLO level, the output of the regulator is turned off.

7.4.2 Stand-by Mode

The internal LDO has a lower enable threshold than the regulator. When V_EN is above 1.2 V and below the precision enable falling threshold (1.8 V typically), the internal LDO regulates the V_CC voltage at 3.2 V. The precision enable circuitry is turned on once V_CC is above the UVLO threshold. The switching action and voltage regulation are not enabled unless V_EN rises above the precision enable threshold (2.1 V typically).

7.4.3 Active Mode

The LM43603-Q1 is in active mode when V_EN is above the precision enable threshold and V_CC is above its UVLO level. The simplest way to enable the LM43603-Q1 is to connect the EN pin to V_IN. This allows self start-up when the input voltage is in the operation range: 3.5 V to 36 V. See Enable (EN) and VCC, UVLO, and BIAS for details on setting these operating levels.

In active mode, depending on the load current, the LM43603-Q1 will be in one of four modes:

1. Continuous conduction mode (CCM) with fixed switching frequency when load current is above half of the peak-to-peak inductor current ripple;
2. Discontinuous conduction mode (DCM) with fixed switching frequency when load current is lower than half of the peak-to-peak inductor current ripple in CCM operation;
3. Pulse frequency modulation (PFM) when switching frequency is decreased at very light load;
4. Fold-back mode when switching frequency is decreased to maintain output regulation at lower supply voltage V_IN.

7.4.4 CCM Mode

CCM operation is employed in the LM43603-Q1 when the load current is higher than half of the peak-to-peak inductor current. In CCM operation, the frequency of operation is fixed by internal oscillator unless the the minimum HS switch ON time (T_{ON_MIN}) or OFF time (T_{OFF_MIN}) is exceeded. Output voltage ripple is at a minimum in this mode and the maximum output current of 2 A can be supplied by the LM43603-Q1.

7.4.5 Light Load Operation

When the load current is lower than half of the peak-to-peak inductor current in CCM, the LM43603-Q1 operates in DCM, also known as diode emulation mode (DEM). In DCM operation, the LS FET is turned off when the inductor current drops to 0 A to improve efficiency. Both switching losses and conduction losses are reduced in DCM, comparing to forced PWM operation at light load.

At even lighter current loads, PFM is activated to maintain high efficiency operation. When the HS switch ON time reduces to T_{ON_MIN} or peak inductor current reduces to its minimum I_PEAK-MIN, the switching frequency reduces to maintain proper regulation. Efficiency is greatly improved by reducing switching and gate drive losses.

Figure 45. V_OUT = 5 V, Fs = 500 kHz
Pulse Frequency Mode Operation

7.4.6 Self-Bias Mode

For highest efficiency of operation,TI recommends that the BIAS pin be connected directly to V_OUT when V_OUT ≥ 3.3 V. In this self-bias mode of operation, the difference between the input and output voltages of the internal LDO are reduced and therefore the total efficiency is improved. These efficiency gains are more evident during light load operation. During this mode of operation, the LM43603-Q1 operates with a minimum quiescent current of 27 µA (typical). See VCC, UVLO, and BIAS for more details.

Table 2. L, C_OUT and C_FF Typical Values

8.2.1 Design Requirements

Detailed design procedure is described based on a design example. For this design example, use the parameters listed in Table 3 as the input parameters.

8.2.2 Detailed Design Procedure

8.2.2.2 Output Voltage Setpoint

The output voltage of the LM43603-Q1 device is externally adjustable using a resistor divider network. The divider network is comprised of top feedback resistor R_FBT and bottom feedback resistor R_FBB. Equation 11 is used to determine the output voltage of the converter:

Equation 11.

Choose the value of the R_FBT to be 100 kΩ to minimize quiescent current to improve light load efficiency in this application. With the desired output voltage set to be 3.3 V and the V_FB = 1.015 V, the R_FBB value can then be calculated using Equation 11. The formula yields a value of 43.478 kΩ. Choose the closest available value of 43.2 kΩ for the R_FBB. See Adjustable Output Voltage for more details.

8.2.2.5 Inductor Selection

The first criterion for selecting an output inductor is the inductance itself. In most buck converters, this value is based on the desired peak-to-peak ripple current, Δi_L, that flows in the inductor along with the DC load current. As with switching frequency, the selection of the inductor is a tradeoff between size and cost. Higher inductance gives lower ripple current and hence lower output voltage ripple with the same output capacitors. Lower inductance could result in smaller, less expensive component. An inductance that gives a ripple current of 20% to 40% of the 3 A at the typical supply voltage is a good starting point. Δi_L = (1/5 to 2/5) × I_OUT. The peak-to-peak inductor current ripple can be found by Equation 13 and the range of inductance can be found by Equation 14 with the typical input voltage used as V_IN.

Equation 13.

Equation 14.

D is the duty cycle of the converter where in a buck converter case it can be approximated as D = V_OUT / V_IN, assuming no loss power conversion. By calculating in terms of amperes, volts, and megahertz, the inductance value will come out in micro Henries. The inductor ripple current ratio is defined by:

Equation 15.

The second criterion is inductor saturation current rating. The inductor must be rated to handle the maximum load current plus the ripple current:

Equation 16. I_L-PEAK = I_LOAD-MAX + Δi_L/ 2

The LM43603-Q1 has both valley current limit and peak current limit. During an instantaneous short, the peak inductor current can be high due to a momentary increase in duty cycle. The inductor current rating must be higher than the HS current limit. It is advised to select an inductor with a larger core saturation margin and preferably a softer roll off of the inductance value over load current.

In general, it is preferable to choose lower inductance in switching power supplies, because it usually corresponds to faster transient response, smaller DCR, and reduced size for more compact designs. But too low of an inductance can generate too large of an inductor current ripple such that over current protection at the full load could be falsely triggered. It also generates more conduction loss, because the RMS current is slightly higher relative that with lower current ripple at the same DC current. Larger inductor current ripple also implies larger output voltage ripple with the same output capacitors. With peak current mode control, it is not recommended to have too small of an inductor current ripple. A larger peak current ripple improves the comparator signal to noise ratio.

Once the inductance is determined, the type of inductor must be selected. Ferrite designs have very low core losses and are preferred at high switching frequencies, so design goals can concentrate on copper loss and preventing saturation. Ferrite core material saturates hard, which means that inductance collapses abruptly when the peak design current is exceeded. The ‘hard’ saturation results in an abrupt increase in inductor ripple current and consequent output voltage ripple. Do not allow the core to saturate!

For the design example, a standard 6.8 μH inductor from Würth Elektronik, Coilcraft, or Vishay can be used for the 3.3 V output with plenty of current rating margin.

8.2.2.6 Output Capacitor Selection

The device is designed to be used with a wide variety of LC filters. Use as little output capacitance as possible to keep cost and size down. Choose the output capacitor(s), C_OUT, with care because it directly affects the steady state output voltage ripple, loop stability and the voltage over/undershoot during load current transients.

The output voltage ripple is essentially composed of two parts. One is caused by the inductor current ripple going through the ESR of the output capacitors:

Equation 17. ΔV_OUT-ESR = Δi_L × ESR

The other is caused by the inductor current ripple charging and discharging the output capacitors:

Equation 18. ΔV_OUT-C =Δi_L/ (8 × F_S × C_OUT )

The two components in the voltage ripple are not in phase, so the actual peak-to-peak ripple is smaller than the sum of the two peaks.

Output capacitance is usually limited by transient performance specifications if the system requires tight voltage regulation with presence of large current steps and fast slew rates. When a fast large load transient happens, output capacitors provide the required charge before the inductor current can slew to the appropriate level. The initial output voltage step is equal to the load current step multiplied by the ESR. V_OUT continues to droop until the control loop response increases or decreases the inductor current to supply the load. To maintain a small overshoot or undershoot during a transient, small ESR and large capacitance are desired. But these also come with higher cost and size. Thus, the motivation is to seek a fast control loop response to reduce the output voltage deviation.

For a given input and output requirement, Equation 19 gives an approximation for an absolute minimum output capacitor required:

Equation 19.

Along with this for the same requirement, calculate the maximum ESR as per Equation 20:

Equation 20.

where

• r = Ripple ratio of the inductor ripple current (ΔI_L / I_OUT)
• ΔV_OUT = target output voltage undershoot
• D’ = 1 – duty cycle
• F_S = switching frequency
• I_OUT = load current

A general guideline for C_OUT range is that C_OUT must be larger than the minimum required output capacitance calculated by Equation 19, and smaller than 10 times the minimum required output capacitance or 1 mF. In applications with V_OUT less than 3.3 V, it is critical that low ESR output capacitors are selected. This limits potential output voltage overshoots as the input voltage falls below the device normal operating range. To optimize the transient behavior a feed-forward capacitor could be added in parallel with the upper feedback resistor. For this design example, three 47-µF, 10-V, X7R ceramic capacitors are used in parallel.

8.2.2.7 Feed-Forward Capacitor

The LM43603-Q1 is internally compensated and the internal R-C values are 400 kΩ and 50 pF, respectively. Depending on the V_OUT and frequency F_S, if the output capacitor C_OUT is dominated by low ESR (ceramic types) capacitors, it could result in low phase margin. To improve the phase boost an external feedforward capacitor C_FF can be added in parallel with R_FBT. C_FF is chosen such that phase margin is boosted at the crossover frequency without C_FF. A simple estimation for the crossover frequency without C_FF (f_x) is shown in Equation 21, assuming C_OUT has very small ESR.

Equation 21.

Equation 22 was tested for C_FF:

Equation 22.

This equation indicates that the crossover frequency is geometrically centered on the zero and pole frequencies caused by the C_FF capacitor.

For designs with higher ESR, C_FF is not needed when C_OUT has very high ESR, and C_FF calculated from Equation 22 must be reduced with medium ESR. Table 2 can be used as a quick starting point.

For the application in this design example, a 470 pF COG capacitor is selected.

8.2.2.11 Soft-Start Capacitors

The user can left the SS/TRK pin floating, and the LM43603-Q1 implements a soft-start time of 4.1 ms typically. In order to use an external soft-start capacitor, the capacitor must be sized so that the soft start time is longer than 4.1 ms. Use Equation 23 in order to calculate the soft start capacitor value:

Equation 23.

where

• C_SS = Soft start capacitor value (µF)
• I_SS = Soft start charging current (µA)
• t_SS = Desired soft start time (s)

For the desired soft-start time of 10 ms and soft-start charging current of 2 µA, Equation 23 above yield a soft start capacitor value of 0.02 µF.

8.2.3 Application Performance Curves

VOUT = 1 V	FS = 500 kHz

Figure 49. Component Values for V_OUT= 1 V,
F_S = 500 kHz

VOUT = 1 V

FS = 500 kHz

Figure 51. Output Voltage Regulation

VIN = 12 V

VOUT = 1 V

Figure 53. Load Transient 0.1 A to 1 A

VOUT = 3.3 V	FS = 500 kHz

Figure 55. Component Values for V_OUT = 3.3 V,
F_S = 500 kHz

VOUT = 3.3 V

FS = 500 kHz

Figure 57. Efficiency at Room Temperature

VOUT = 3.3 V

FS = 500 kHz

Figure 59. Power Loss at Room Temperature

VOUT = 3.3 V

FS = 500 kHz

Figure 61. Dropout Curve

VOUT = 3.3 V

FS = 500 kHz

Figure 63. Frequency vs Load

VOUT = 3.3 V

FS = 500 kHz

Figure 65. Load Transient 0.1 A to 2 A

VOUT = 5 V	FS = 200 kHz

Figure 67. Component Values for V_OUT = 5 V,
F_S = 200 kHz

VOUT = 5 V

FS = 200 kHz

Figure 69. Output Voltage Regulation

VOUT = 5 V

FS = 200 kHz

Figure 71. Load Transient 0.1 A to 2 A

VOUT = 5 V	FS = 500 kHz

Figure 73. Component Values for V_OUT = 5 V,
F_S = 500 kHz

VOUT = 5 V

FS = 500 kHz

Figure 75. Output Voltage Regulation

VOUT = 5 V

FS = 500 kHz

Figure 77. Load Transient 0.1 A to 2 A

VOUT = 5 V	FS = 1 MHz

Figure 79. Component Values for V_OUT = 5 V,
F_S = 1 MHz

VOUT = 5 V

FS = 1 MHz

Figure 81. Output Voltage Regulation

VOUT = 5 V

FS = 1 MHz

Figure 83. Load Transient

VOUT = 5 V	FS = 2.2 MHz

Figure 85. Component Values for V_OUT = 5 V,
F_S = 2.2 MHz

VOUT = 5 V

FS = 2.2 MHz

Figure 87. Output Voltage Regulation

VOUT = 5 V

FS = 2.2 MHz

Figure 89. Load Transient

VOUT = 12 V	FS = 500 kHz

Figure 91. Component Values for V_OUT = 12 V,
F_S = 500 kHz

VOUT = 12 V

FS = 500 kHz

Figure 93. Output Voltage Regulation

VOUT = 12 V

FS = 500 kHz

VIN = 24 V

Figure 95. Load Transient 0.1 A to 2 A

VOUT = 1 V

Fs = 500 kHz

Figure 50. Efficiency

VOUT = 1 V

FS = 500 kHz

Figure 52. Frequency vs Load

VOUT = 1 V

FS = 500 kHz

RθJA = 20°C/W

Figure 54. Derating Curve

VOUT = 3.3 V

FS = 500 kHz

Figure 56. Efficiency at Room Temperature

VOUT = 3.3 V

FS = 500 kHz

Figure 58. Efficiency at 85ºC Ambient Temperature

VOUT = 3.3 V

FS = 500 kHz

Figure 60. Power Loss at 85°C Ambient Temperature

VOUT = 3.3 V

FS = 500 kHz

Figure 62. Frequency vs V_IN

VOUT = 3.3 V

FS = 500 kHz

Figure 64. Output Voltage Regulation

VOUT = 3.3 V

FS = 500 kHz

RθJA = 20°C/W

Figure 66. Derating Curve

VOUT = 5 V

FS = 200 kHz

Figure 68. Efficiency at Room Temperature

VOUT = 5 V

FS = 200 kHz

Figure 70. Dropout Curve

VOUT = 5 V

FS = 200 kHz

RθJA = 20°C/W

Figure 72. Derating Curve

VOUT = 5 V

FS = 500 kHz

Figure 74. Efficiency at Room Temperature

VOUT = 5 V

FS = 500 kHz

Figure 76. Dropout Curve

VOUT = 5 V

FS = 500 kHz

RθJA = 20°C/W

Figure 78. Derating Curve

VOUT = 5 V

FS = 1 MHz

Figure 80. Efficiency

VOUT = 5 V

FS = 1 MHz

Figure 82. Dropout Curve

VOUT = 5 V

FS = 1 MHz

RθJA = 20°C/W

Figure 84. Derating Curve

VOUT = 5 V

FS = 2.2 MHz

Figure 86. Efficiency

VOUT = 5 V

FS = 2.2 MHz

Figure 88. Dropout Curve

VOUT = 5 V

FS = 2.2 MHz

RθJA = 20°C/W

Figure 90. Derating Curve

VOUT = 12 V

FS = 500 kHz

Figure 92. Efficiency

VOUT = 12 V

FS = 500 kHz

Figure 94. Dropout Curve

VOUT = 12 V

FS = 500 kHz

RθJA = 20°C/W

Figure 96. Derating Curve