Switched Capacitor Regulator
Time：2021.10.26
View：
The growing trend in today's mobile phone industry is to reduce the voltage supply to the core processor while meeting the requirements for achieving higher efficiency and extending battery life. More and more applications require step-down conversion, such as application processors, memory and RF block designs are listed among them. According to load and space parameters, currently, the two most popular solutions for this application space are switching regulators and low dropout (LDO) regulators.
From an efficiency point of view, a switching regulator is the best choice. However, when the component height and the size limit of the solution exceed the range of use of the inductor, a converter may take the form of an LDO (low dropout) or switched capacitor regulator. Most of the time power solutions can not provide more board space, a switching converter will have a larger solution size than LDO and switched capacitor regulator. Figure 1 compares the solution size between a typical switching regulator and the LM2770 (a typical switched capacitor regulator). We can see that the displayed switched capacitor solution size is about 45mm2. When the required voltage is close to the battery voltage, the LDO's working efficiency is the highest, but when the voltage deviates very far, the LDO efficiency will be very high. Low. Imagine using a Li-Ion battery charged to 3.6V to power a microprocessor that only requires 1.5V. Connecting the battery voltage to a 1.5V LDO can generate a stable and small current for the microprocessor, but the power consumption is quite significant. The LDO power consumption (PD) is equal to the load current (ILOAD) multiplied by the difference between the input and output voltages (PD = ILOAD *(3.6-1.5) = ILOAD *2.3V). In other words, the LDO as a buck converter produces only 42% efficiency in this example. This means that the LDO has to consume surplus power, and this can cause a large increase in die temperature, which in turn may cause reliability problems.
Due to the voltage gain, a switched capacitor regulator is a more effective solution than a linear regulator. This voltage gain is achieved by stacking capacitors and parallel capacitors in the dual phase (charging phase and transmission phase) The ratio of input voltage to output voltage. For example: 1/2 of a switched capacitor converter in a gain configuration will convert an input voltage (VIN) of 3.6V into an output voltage (VOUT) of 1.8V. If the required voltage (VOUT) is 1.5V, the power consumption is only the product of 300mV and the load current. This corresponds to an efficiency of 83%.
As VIN increases, the incremental increase between VIN and VOUT generated by the converter causes an increase in power consumption and a decrease in efficiency. One way to solve this problem is to transform into a higher efficiency gain, just like replacing gears in a car. Figure 2 shows the efficiency curve of a switched capacitor buck regulator, an LDO and a switched capacitor. Switched capacitors have an analog gain control and gain changes to maintain the continuity of a given load efficiency. Switched capacitors also have discrete gain steps, and the efficiency is determined by VOUT/(gain *VIN) and discrete gain. An LDO has only 1 gain and the lowest efficiency of the three. Switched capacitor (SC) regulators have three different voltage gains (2/3, 1/2 and 1/3). We can see that with the increase of VIN, the voltage gain of the switched capacitor regulator changes from 2/3 to 1/2 and from 1/2 to 1/3, so the efficiency of the entire load range is maximized. This brings about 80% efficiency in the Li-ion range (3.4V to 3.8V). An LDO in the same application can only achieve 50% efficiency. Depending on the type of inductor, a typical switching regulator has an efficiency of approximately 88-90%.
Flying capacitor or CFLY. CIN and COUT have been deleted for simplification purposes. As shown in the figure, a gain is obtained by alternating changes between two phases, including charging phase or normal phase and discharging phase. There is a common phase between different gains in order to achieve a perfect transition between gains. We can make gain transitions at any time as needed through the common phase. A switched capacitor regulator may have one to two power FETs on the chip. However, a switched capacitor regulator may have 4 to 9 or more power FETs (depending on the number of discrete voltage gains) anywhere on the chip. This limits the output current performance of the switched capacitor regulator under a given die size. Figure 4 compares the load performance, efficiency, and size of switches, switched capacitors, and linear regulators.
To use a switched capacitor regulator to regulate the output voltage, we can use pulse frequency modulation (PFM) or pulse width modulation (PWM). The output impedance of any switched capacitor regulator is proportional to the switching frequency and the resistance of the internal power FET. By modulating the output impedance, we can use the converter to step down a given load. By using feedback, we can control the frequency (PFM) or the impedance of the internal FET (PWM) to adjust the output voltage. The PFM scheme is a more traditional approach, and its shortcomings are listed in recent PWM-type architectures.
In a PFM system, the output voltage can be sensed. When this voltage is higher than a specified value, the voltage regulator will be turned off and will be turned on again when the output voltage drops below the required value. The disadvantage of using the PFM control mode is that the operating frequency depends on VIN and ILOAD, so it varies. The higher the load, the closer the operating frequency is to the specified frequency. The frequency change over this operating range may not be appropriate in some portable applications. The input voltage ripple also depends on VIN and ILOAD. Figure 5 shows the output ripple for 250mA and 30mA loads. The output ripple of 10mF COUT will be 50mV, we can see that the ripple frequency of 250mA load is higher than that of 10mA load.
The relatively new PWM control mode handles various frequencies and high output ripples in the PFM architecture. Most of the new switched capacitor regulators use PWM modulation mode. In this mode, the resistance of the power FET is controlled according to VOUT and ILOAD. In doing so, we truly control the charging capacity provided by the flying capacitor (CFLY). This is also called pre-modulation. In this mode, the operating frequency and duty cycle are fixed. An example of a PWM architecture is the LM2771, and Figure 6 shows its output ripple. It is in a sequence of 8mV -10mV with 4.7mF COUT. We can see that the ripple can be sustained when ILOAD changes. A 9mV ripple output is comparable to the ripple in an inductive switching regulator.
Switched capacitor regulator is an emerging technology that combines the main characteristics of switched capacitor and LDO, that is, combining high efficiency and small size in the Li-Ion range into a simple solution suitable for portable applications. The recent development of topology technology also enables it to achieve lower noise through smaller values of passive components. Many functions in portable devices require a buck regulator, smaller solution size and higher efficiency. This switched capacitor regulator solution is an ideal solution.
