Table 1. Bill of Materials for Figure 30

8.2.1.2.1 Inductor Selection

The duty cycle (D) can be approximated quickly using the ratio of output voltage (V_O) to input voltage (V_IN), using Equation 11:

Equation 11.

The catch diode (D1) forward voltage drop and the voltage drop across the internal NMOS switch must be included to calculate a more accurate duty cycle. Calculate D by using Equation 12:

Equation 12.

V_SW can be approximated by Equation 13:

The diode forward drop (V_D) can range from 0.3 V to 0.7 V depending on the quality of the diode. The lower the V_D, the higher the operating efficiency of the converter. The inductor value determines the output ripple current. Lower inductor values decrease the size of the inductor, but increase the output ripple current. An increase in the inductor value decreases the output ripple current.

One must ensure that the minimum current limit (2 A) is not exceeded, so the peak current in the inductor must be calculated. The peak current (I_LPK) in the inductor is calculated by Equation 14 and Equation 15:

Figure 31. Inductor Current

Equation 15.

In general in Equation 16,

If Δi_L = 33.3% of 1.5 A, the peak current in the inductor is 2 A. The minimum specified current limit over all operating conditions is 2 A. One can either reduce Δi_L, or make the engineering judgment that zero margin is safe enough. The typical current limit is 2.9 A.

The LM2738 operates at frequencies allowing the use of ceramic output capacitors without compromising transient response. Ceramic capacitors allow higher inductor ripple without significantly increasing output ripple. See the Output Capacitor section for more details on calculating output voltage ripple. Now that the ripple current is determined, the inductance is calculated by Equation 17:

Equation 17.

where

• 

When selecting an inductor, make sure that it is capable of supporting the peak output current without saturating. Inductor saturation results in a sudden reduction in inductance and prevents the regulator from operating correctly. Because of the speed of the internal current limit, the peak current of the inductor need only be specified for the required maximum output current. For example, if the designed maximum output current is 1 A and the peak current is 1.25 A, the inductor must be specified with a saturation current limit of > 1.25 A. There is no must specify the saturation or peak current of the inductor at the 2.9-A typical switch current limit. Because of the operating frequency of the LM2738, ferrite based inductors are preferred to minimize core losses. This presents little restriction because of the variety of ferrite-based inductors available. Lastly, inductors with lower series resistance (R_DCR) provide better operating efficiency. For recommended inductors see LM2738X Circuit Example 1.

8.2.1.2.2 Input Capacitor

An input capacitor is necessary to ensure that V_IN does not drop excessively during switching transients. The primary specifications of the input capacitor are capacitance, voltage, RMS current rating, and equivalent series inductance (ESL). The recommended input capacitance is 10 µF. The input voltage rating is specifically stated by the capacitor manufacturer. Make sure to check any recommended deratings and also verify if there is any significant change in capacitance at the operating input voltage and the operating temperature. The input capacitor maximum RMS input current rating (I_RMS-IN) must be greater than Equation 18:

Equation 18.

Neglecting inductor ripple simplifies Equation 18 to Equation 19:

Equation 19.

Equation 19 shows that maximum RMS capacitor current occurs when D = 0.5. Always calculate the RMS at the point where the duty cycle D is closest to 0.5. The ESL of an input capacitor is usually determined by the effective cross-sectional area of the current path. A large leaded capacitor has high ESL and a 0805 ceramic-chip capacitor has very low ESL. At the operating frequencies of the LM2738, leaded capacitors may have an ESL so large that the resulting impedance (2πfL) is higher than that required to provide stable operation. As a result, surface-mount capacitors are strongly recommended.

Sanyo POSCAP, Tantalum or Niobium, Panasonic SP, and multilayer ceramic capacitors (MLCC) are all good choices for both input and output capacitors and have very low ESL. For MLCCs, TI recommends using X7R or X5R type capacitors due to their tolerance and temperature characteristics. Consult the capacitor manufacturer's data sheets to see how rated capacitance varies over operating conditions.

8.2.1.2.3 Output Capacitor

The output capacitor is selected based upon the desired output ripple and transient response. The initial current of a load transient is provided mainly by the output capacitor. The output ripple of the converter is Equation 20:

Equation 20.

When using MLCCs, the equivalent series resistance (ESR) is typically so low that the capacitive ripple may dominate. When this occurs, the output ripple is approximately sinusoidal and 90° phase shifted from the switching action. Given the availability and quality of MLCCs and the expected output voltage of designs using the LM2738, there is really no must review any other capacitor technologies. Another benefit of ceramic capacitors is the ability to bypass high-frequency noise. A certain amount of switching edge noise couples through parasitic capacitances in the inductor to the output. A ceramic capacitor bypasses this noise while a tantalum capacitor does not. Since the output capacitor is one of the two external components that control the stability of the regulator control loop, most applications require a minimum of 22 µF of output capacitance. Capacitance, in general, is often increased when operating at lower duty cycles. Refer to the Circuit Examples for suggested output capacitances of common applications. Like the input capacitor, recommended multilayer ceramic capacitors are X7R or X5R types.

8.2.1.2.4 Catch Diode

The catch diode (D1) conducts during the switch off time. A Schottky diode is recommended for its fast switching times and low forward voltage drop. The catch diode must be chosen so that its current rating is greater than Equation 21:

The reverse breakdown rating of the diode must be at least the maximum input voltage plus appropriate margin. To improve efficiency, choose a Schottky diode with a low forward-voltage drop.

8.2.1.2.5 Output Voltage

The output voltage is set using Equation 22 and Equation 23 where R2 is connected between the FB pin and GND, and R1 is connected between V_O and the FB pin. A good value for R2 is 10 kΩ. When designing a unity gain converter (V_O = 0.8 V), R1 must be between 0 Ω and 100 Ω, and R2 must not be loaded.

Equation 22.

8.2.1.2.6 Calculating Efficiency and Junction Temperature

The complete LM2738 DC-DC converter efficiency can be calculated by Equation 24 or Equation 25:

Equation 24.

or,

Equation 25.

Calculations for determining the most significant power losses are shown in Equation 26. Other losses totaling less than 2% are not discussed.

Power loss (P_LOSS) is the sum of two basic types of losses in the converter: switching and conduction. Conduction losses usually dominate at higher output loads, whereas switching losses remain relatively fixed and dominate at lower output loads. The first step in determining the losses is to calculate the duty cycle (D):

Equation 26.

V_SW is the voltage drop across the internal NFET when it is on, and is equal to Equation 27:

V_D is the forward voltage drop across the Schottky catch diode. It can be obtained from the diode manufacturer's data sheet Electrical Characteristics section. If the voltage drop across the inductor (V_DCR) is accounted for, the equation becomes Equation 28:

Equation 28.

The conduction losses in the free-wheeling Schottky diode are calculated by Equation 29:

Often this is the single most significant power loss in the circuit. Care must be taken to choose a Schottky diode that has a low forward-voltage drop.

Another significant external power loss is the conduction loss in the output inductor. The equation can be simplified to Equation 30:

The LM2738 conduction loss is mainly associated with the internal NFET switch in Equation 31:

Equation 31.

If the inductor ripple current is fairly small, the conduction losses can be simplified to Equation 32:

Switching losses are also associated with the internal NFET switch. They occur during the switch on and off transition periods, where voltages and currents overlap resulting in power loss. The simplest means to determine this loss is to empirically measure the rise and fall times (10% to 90%) of the switch at the switch node.

Switching Power Loss is calculated as follows in Equation 33, Equation 34, and Equation 35:

Another loss is the power required for operation of the internal circuitry in Equation 36:

I_Q is the quiescent operating current, and is typically around 1.9 mA for the 0.55-MHz frequency option.

Table 2 lists the power losses for a typical application, and in Equation 37, Equation 38, and Equation 39.

Table 3. Bill of Materials for Figure 33

Table 4. Bill of Materials for Figure 35

Table 5. Bill of Materials for Figure 37

Table 6. Bill of Materials for Figure 38

Table 7. Bill of Materials for Figure 39

Table 8. Bill of Materials for Figure 41

Table 9. Bill of Materials for Figure 43

Table 10. Bill of Materials for Figure 45

Table 11. Bill of Materials for Figure 47