8.2.2.1 Input Filter Capacitor

Fast switching currents place a large strain on the input supply to a buck regulator. A capacitor placed close to the PVIN and PGND pins of the LM2854 helps to supply the instantaneous charge required when the regulator demands a pulse of current every switching cycle. In fact, the input capacitor conducts a square-wave current of peak-to-peak amplitude equal to I_OUT. With this high AC current present in the input capacitor, the RMS current rating becomes an important parameter. The necessary RMS current rating of the input capacitor to a buck regulator can be estimated using Equation 1.

Equation 1.

where the PWM duty cycle, D, is given in Equation 2.

Equation 2.

Neglecting capacitor ESR, the resultant input capacitor AC ripple voltage is a triangular waveform with peak-to-peak amplitude specified in Equation 3.

Equation 3.

The maximum input capacitor ripple voltage and RMS current occur at 50% duty cycle. A 22-µF or 47-µF high-quality dielectric (X5R, X7R) ceramic capacitor with adequate voltage rating is typically sufficient as an input capacitor to the LM2854. The input capacitor should be placed as close as possible to the PVIN and PGND pins to substantially eliminate the parasitic effects of any stray inductance or resistance on the PC board and supply lines. Additional bulk capacitance with higher ESR may be required to damp any resonance effects of the input capacitance and parasitic inductance.

8.2.2.2 AVIN Filtering Components

In addition to the large input filter capacitor, a smaller ceramic capacitor such as a 0.1 µF or 1.0 µF is recommended between AVIN and AGND to filter high frequency noise present on the PVIN rail from the quiet AVIN supply. For additional filtering in noisy environments, a small RC filter can be used on the AVIN pin as shown below.

Figure 16. Filtering of AVIN

In general, R_F is typically selected between 1 Ω and 10 Ω so that the steady state voltage drop across the resistor due to the AVIN bias current does not affect the UVLO level. Recommended filter capacitor, C_F, is 1 µF in X5R or X7R dielectric.

8.2.2.3 Soft-Start Capacitor

When the LM2854 is enabled, the output voltage will ramp up linearly in the time dictated by the relationship shown in Equation 4.

Equation 4.

where V_REF is the internal reference voltage (nominally 0.8V), I_SS is the soft-start charging current (nominally 2 µA) and C_SS is the external soft-start capacitance. Rearranging this equation allows for the necessary soft-start capacitor for a given start-up time to be calculated as in Equation 5.

Equation 5.

Thus, the required soft start capacitor per unit output voltage start-up time is given in Equation 6.

For example, a 10 nF soft-start capacitor will yield a 4 ms soft-start time.

8.2.2.4 Tracking - Equal Soft-Start Time

One way to use the tracking feature is to design the tracking resistor divider so that the master supply output voltage, V_OUT1, and the LM2854 output voltage, V_OUT2, both rise together and reach their target values at the same time. This is termed ratiometric start-up. For this case, the equation governing the values of tracking divider resistors R_T1 and R_T2 is given in Equation 7.

Equation 7.

The above equation includes an offset voltage to ensure that the final value of the SS pin voltage exceeds the reference voltage of the LM2854. This offset will cause the LM2854 output voltage to reach regulation slightly before the master supply. A value of 33 kΩ 1% is recommended for R_T2 as a compromise between high precision and low quiescent current through the divider while minimizing the effect of the 2 µA soft-start current source.

For example, If the master supply voltage V_OUT1 is 3.3V and the LM2854 output voltage was 1.8V, then the value of R_T1 needed to give the two supplies identical soft-start times would be 14.3 kΩ. A timing diagram for this example, the equal soft-start time case, is shown in Figure 17.

Figure 17. Simplified Start-up Waveforms When Using Proportional Tracking

8.2.2.5 Tracking - Equal Slew Rates

Alternatively, the tracking feature can be used to have similar output voltage ramp rates. This is referred to as simultaneous start-up. In this case, the tracking resistors can be calculated using Equation 8.

Equation 8.

and to ensure proper overdrive of the SS pin as calculated in Equation 9.

For the example case of V_OUT1 = 5 V and V_OUT2 = 2.5 V, with RT2 set to 33 kΩ as before, R_T1 is calculated from the above equation to be 15.5 kΩ. A timing diagram for the case of equal slew rates is shown in Figure 18.

Figure 18. Simplified Start-up Waveforms, Showing Tracking Used to Achieve Equal Slew Rates

8.2.2.6 Enable and UVLO

Using a resistor divider from VIN to EN as shown in the schematic diagram below, the input voltage at which the part begins switching can be increased above the normal input UVLO level as shown in Equation 10.

Equation 10.

For example, suppose that the required input UVLO level is 3.69 V. Choosing R_EN2 = 10 kΩ, then we calculate R_EN1 = 20 kΩ.

Figure 19. Simplified Schematic Showing Use of EN as an Input UVLO

Alternatively, the EN pin can be driven from another voltage source to cater for system sequencing requirements commonly found in FPGA and other multi-rail applications. The following schematic shows an LM2854 that is sequenced to start based on the voltage level of a master system rail.

Figure 20. Simplified Schematic Showing EN Used to Cascade Power Supply Start-up

8.2.2.7 Output Voltage Setting

A divider resistor network from V_OUT to the FB pin determines the desired output voltage as shown in Equation 11.

Equation 11.

R_FB1 is defined based on the voltage loop requirements and R_FB2 is then selected for the desired output voltage. These resistors are normally selected as 0.5% or 1% tolerance.

8.2.2.8 Compensation Component Selection

The power stage transfer function of a voltage mode buck converter has a complex double pole related to the LC output filter and a left half plane zero due to the output capacitor ESR, denoted R_ESR. The locations of these singularities are given respectively in Equation 12.

Equation 12.

where

• C_O is the output capacitance value appropriately derated for applied voltage and operating temperature
• R_L is the effective load resistance
• R_DCR is the series damping resistance associated with the inductor and power switches

Figure 21. LM2854 Compensation Scheme

The conventional compensation strategy employed with voltage mode control is to use two compensator zeros to offset the LC double pole, one compensator pole located to cancel the output capacitor ESR zero and one compensator pole located between one third and one half switching frequency for high frequency noise attenuation.

The LM2854 internal compensation components are designed to locate a pole at the origin and a pole at high frequency as mentioned above. Furthermore, a zero is located at 8.8 kHz or 17.6 kHz for the 500 kHz or 1 MHz options, respectively, to approximately cancel the likely location of one LC filter pole.

The three external compensation components, R_FB1, R_COMP and C_COMP, are selected to position a zero at or below the LC pole location and a pole to cancel the ESR zero. The voltage loop crossover frequency, f_loop, is usually selected between one tenth to one fifth of the switching frequency, as shown in Equation 13.

A simple solution for the required external compensation capacitor, C_COMP, with type III voltage mode control can be expressed as in Equation 14.

Equation 14.

where the constant α is nominally 0.038 or 0.075 for the 500 kHz or 1 MHz options, respectively. This assumes a compensator pole cancels the output capacitor ESR zero. Furthermore, since the modulator gain is proportional to V_IN, the loop crossover frequency increases with V_IN. Thus, it is recommended to design the loop at maximum expected V_IN.

The upper feedback resistor, R_FB1, is selected to provide adequate mid-band gain and to locate a zero at or below the LC pole frequency. The series resistor, R_COMP, is selected to locate a pole at the ESR zero frequency, as shown in Equation 15.

Equation 15.

Note that the lower feedback resistor, R_FB2, has no impact on the control loop from an AC standpoint since the FB pin is the input to an error amplifier and effectively at AC ground. Hence, the control loop can be designed irrespective of output voltage level. The only caveat here is the necessary derating of the output capacitance with applied voltage. Having chosen R_FB1 as above, R_FB2 is then selected for the desired output voltage.

Table 1 and Table 2 list inductor and ranges of capacitor values that work well with the LM2854, along with the associated compensation components to ensure stable operation. Values different than those listed may be used, but the compensation components may need to be recalculated to avoid degradation in phase margin. Note that the capacitance ranges specified refer to in-circuit values where the nominal capacitance value is adequately derated for applied voltage.

8.2.2.9 Filter Inductor and Output Capacitor Selection

In a buck regulator, selection of the filter inductor and capacitor will affect many key system parameters, including stability, transient response and efficiency The LM2854 can accommodate relatively wide ranges of output capacitor and filter inductor values in a typical application and still achieve excellent load current transient performance and low output voltage ripple.

The inductance is chosen such that the peak-to-peak inductor current ripple, Δi_L, is approximately 25 to 40% of I_OUT as shown in Equation 16.

Equation 16.

Note that the peak inductor current is the DC output current plus half the ripple current and reaches its highest level at lowest duty cycle (or highest V_IN). It is recommended that the inductor should have a saturation current rating in excess of the current limit level.

When operating the LM2854 at input voltages above 5.2 V, the inductor should be sized to keep the minimum inductor current above –0.5 A. For most applications this should only occur at light loads or when the inductor is drastically undersized. To ensure the current never goes below –0.5 A for any application, the peak-to-peak ripple current (Δi_L) in the inductor should be less than 1 A. Keeping the minimum inductor current above –0.5 A limits the energy storage in the inductor and helps prevent the switch node voltage from exceeding the absolute maximum specification when the low side FET turns off.

Table 3 lists examples of off-the-shelf powdered iron and ferrite based inductors that are suitable for use with the LM2854. The output capacitor can be of ceramic or electrolytic chemistry. The chosen output capacitor requires sufficient DC voltage rating and RMS ripple current handling capability.

The output capacitor RMS current and peak-to-peak output ripple are given respectively as in Equation 17.

Equation 17.

In general, 22 µF to 100 µF of ceramic output capacitance is sufficient for both LM2854 frequency options given the optimal high frequency characteristics and low ESR of ceramic dielectric. It is advisable to consult the manufacturer’s derating curves for capacitance voltage coefficient as the in-circuit capacitance may drop significantly with applied voltage.

Tantalum or organic polymer electrolytic capacitance may be suitable with the LM2854 500 kHz option, particularly in applications where substantial bulk capacitance per unit volume is required. However, the high loop bandwidth achievable with the LM2854 obviates the necessity for large bulk capacitance during transient loading conditions.

Table 4 lists some examples of commercially available capacitors that can be used with the LM2854.

Table 5. LM2854 500-kHz Bill of Materials, V_IN = 5 V, V_OUT = 3.3 V, I_OUT(MAX) = 4 A, Optimized for Efficiency

Table 6. LM2854 1-MHz Bill of Materials, V_IN = 3.3 V to 5 V, V_OUT = 2.5 V, I_{OUT (MAX)} = 4 A, Optimized for Electrolytic Input and Output Capacitance

Table 7. LM2854 1-MHz Bill of Materials, V_IN = 3.3 V, V_OUT = 0.8 V, I_{OUT (MAX)} = 4 A, Optimized for Solution Size and Transient Response