Power supplies 101

Power supplies are essential components of electronic devices. This article details the basics of linear and switching power supplies, along with their benefits.

In the most basic terms, a step-down regulator takes an input voltage and converts it to a lower voltage required by the load. There are two major types of voltage regulators: linear and switching.

A voltage regulator that is placed close to the device or devices for which it is powering is commonly referred to as a point of load regulator. This type of regulator often receives its input from a voltage bus contained on the PCB.

A linear regulator works by absorbing the energy difference between the input and the output while a switching regulator chops the input voltage and then averages the chopped voltage to provide the proper output voltage.

There are several parameters that should be considered when evaluating power supplies for particular applications:

Quiescent current: The amount of current an IC consumes when it is not source current and its inputs are not cycling.

Efficiency: Defined as the ratio of power into a regulator circuit to the output power delivered by the regulator. Power input includes quiescent currents, losses due to parasitic components and power delivered to the load.

Equivalent series resistance (ESR): The resistive component of the capacitors used in a regulator circuit. Attention needs to be paid to the ESR of the capacitors of the output.

The loop: Defined as all the elements affecting the output and stability of a regulator. This includes such blocks as the power stage, error amplifier, modulator and compensation circuits.

Additionally, as illustrated in Figure 1, several other specifications associated with power supplies affect their operation. These include:

T_J-MAX — The highest allowable temperature the junction in the IC is allowed to achieve.

T_A — Defined as the ambient temperature of the air in the vicinity of the regulator.

R_THETA-JA — The thermal resistance from the junction to ambient. This includes the thermal resistance from the junction to case, the case to PCB and PCB to air.

V_FB — The feedback voltage used to set the output voltage. R1 and R2 create a resistor divider to form V_FB from V_OUT. The error amp works to keep V_FB equal to V_REF.

Figure 1: Voltage regulator definitions.

Linear regulators are often used because of their ease of use and typically have a very low component count requirement. This type of regulator is equivalent to an adjustable resistor which drops V_IN to V_OUT and tries to keep V_OUT constant. The following is a list of key parameters to look for in a linear regulator:

Dropout voltage — This is the minimum voltage needed above V_OUT to maintain proper regulation.

Low drop out (LDO) — This is a type of regulator that functions when V_DROPOUT is very small.

Stability region — The area of a linear regulator where the output will remain stable. It is given for different output capacitances and depends on the ESR of the output capacitor.

P_D-MAX — The maximum power a device can dissipate is defined as the difference of T_J-MAX and T_AMBIENT divided by R_THETA-JA. It is possible to figure out what the maximum allowable output current is by combining and rearranging the power dissipation equations.

Linear regulators can be classified as either integrated or discrete (as illustrated in Figure 2). An integrated linear regulator has the pass element integrated in the package. A discrete solution has an external pass element. Feedback components may or may not be integrated for either a discrete or integrated regulator. Semtech’s SC339 is an example of a discrete linear regulator. The Semtech SC4215H is an example of an integrated linear regulator.

Figure 2: Integrated vs discrete linear regulators.

In the example shown in Figure 3, the V_IN has been defined to be 3.3 V and V_OUT has been defined as 2.5 V. We begin by selecting R₂ to be 10 kΩ. From the equation for calculating R₁ and using the value of 0.5 V for V_REF from the datasheet, we come up with a value of 40 kΩ needed for R₁. We have defined the T_AMBIENT to be 50°C. From the datasheet, we find the value for T_J-MAX defined as 125°C and R_THETA-JA defined as 36°C/W. With those values, we can calculate the maximum available output current, which is 2.6 A.

Figure 3: Linear regulator example — Semtech SC4216H.

The datasheet states 10 μF per amp of output current is needed to prevent the input to the regulator from sagging. To support the full current calculated on the previous slide, a minimum of 26 μF is needed. Due to capacitor tolerances, a minimum of 33 μF capacitor should be chosen. In addition to the bulk cap, a 4.7 μF ceramic is recommended to be placed directly next to the V_IN pin of the regulator. For the output, the same rule of 10 μF per amp of output current is recommended. This will lead to the same value of 33 μF being used for the output. It is important to watch the ESR of the output capacitor to ensure the regions of instability are not entered. While a graph is not given for a 33 μF capacitor, we can interpolate between the two graphs given and use a 5 mΩ ESR to maintain stability. It is also recommended to use a 0.1 μF capacitor next to the V_OUT pin.

A buck regulator converts a higher V_IN to a lower V_OUT as shown in Figure 4. The controller opens and closes switch S periodically to connect the phase node between V_IN and ground. The waveform seen at the phase node is then filtered through the inductor and capacitor to provide a smooth voltage to the load.

Figure 4: Buck regulator.

The buck regulator has two different states (as shown in Figure 5) when switch S is closed and when switch S is open. When the switch is closed, V_IN charges the inductor and capacitor and supplies the output current. When the switch is open, the magnetic field in the inductor collapses and causes the voltage across the inductor to change polarity. This change in polarity pulls the phase node below ground, forward biasing the diode. This allows current to continue to flow as the inductor sources the current needed by the load.

Figure 5: Basic buck regulator states.

As mentioned previously, the diode becomes forward biased with the collapsing of the magnetic field of the inductor. This occurs without any reference to a clock and is known as an asynchronous regulator. The VF of the diode used represents a significant portion of losses within an asynchronous regulator, especially as the value of V_F approaches the value of V_OUT. By replacing the diode with another MOSFET, the losses in the system can be reduced. This requires a more complex controller to ensure the MOSFETs are not both turned on at the same time. Since the bottom MOSFET is now controlled, the topology is referred to as a synchronous regulator. S₁ is also known as the top FET or the high-side FET. S₂ is also known as the bottom FET or the low-side FET. The comparison between the asynchronous and synchronous buck regulators is shown in Figure 6.

Figure 6: Asynchronous vs synchronous buck regulators.

The waveforms on Figure 7 are associated with a buck regulator. DH is the control signal to the high-side FET. As the controller turns on the high-side FET, current ramps up in the switch. At the same time, current also ramps up in the inductor and the voltage of the phase node hits V_IN. When the high-side switch is turned off, current ramps down in both the inductor and the low-side FET and the voltage at the phase node is pulled to ground.

The signal at V_PHASE is then filtered to form V_OUT.

The amount of time the high-side FET is on is defined as t_ON. The amount of time the high-side FET is off is defined as t_OFF. The time of t_ON plus t_OFF is defined as T or total cycle time.

Figure 7: Basic buck regulator waveforms.

Buck regulators can be further divided into two classes (as depicted in Figure 8): integrated regulators and switching controllers. The integrated switching regulator builds the FETs into the package or in the silicon. The switching controller requires switches to be placed external to the IC.

Figure 8: Integrated switching regulator vs switching controller.

Voltages other than the fixed options provided can be used so long as a feedback network is added to the design. It is recommended to use the 1 V regulator when requiring voltages other than the standard fixed voltages. Due to the internal compensation, the use of a capacitor across the top side of the feedback network is necessary to maintain stability. The equation for calculating the capacitor was already discussed. V_OUT is the desired output voltage and V_OSTD is the voltage of Semtech SC189 with no feedback network. The feedback resistors are calculated via the formula found in Figure 9. The VREF to be used in the equation is the V_OSTD of the part or 1 V when the SC189A is used.

Figure 9: Feedback resistors calculation formula.

In summary, the best time to use a linear regulator is when noise in the system cannot be tolerated or when the V_IN and V_OUT difference is small. The best time to use a switching regulator is when the difference between V_IN and V_OUT is large. Also, it is best to use a switching regulator when efficiency is important and large currents are involved. Semtech has many options for both linear regulators and switching regulators.

For a full audio version of this training, visit: Power Supplies 101.

Top image credit: stock.xchange@xnunox.