Previously it was explained that feedback control for switching regulators comprises three types of control: voltage mode, current mode, and hysteresis control. As noted above, similar to the linear regulator, the switching regulator also performs regulation by using a feedback loop. In this section, details on these control types will be explained. Because each type has its own pros and cons, a particular method must be selected by carefully weighing those factors.
Voltage mode control represents the most basic method, in which only the output voltage is returned through a feedback loop. The differential voltage, which is obtained to compare the output voltage with the reference voltage by an error amp, is compared with triangular waves by a PWM generator. As a result, the pulse width of the PWM signal is determined to control the output voltage. Advantages of this method are its relative simplicity based on the use of a feedback loop consisting solely of voltages, the ability to control shorter on-time, and high noise tolerance. Possible drawbacks are the complexity of the phase compensation circuit and a cumbersome design process.
The current mode is a modification of voltage mode control, where the inductor current in the circuit is detected and used instead of the triangular waveforms used in the voltage mode control. The current sensing can also be done by using the on-resistance of high side MOSFET or a current sense resistor instead of the inductor current. Since the current mode has two types of feedback loops: voltage loop and current loop, the control exerted is relatively complex. However the current mode provides the advantage of a substantially simplified phase compensation circuit design. Other benefits include the highly stable feedback loop and a faster load transient response than that of the voltage mode. A drawback is low-noise tolerance due to the high sensitivity of current detection. In the newer designs, however, the current detection part is built into the IC to alleviate the problem.
Hysteresis (ripple) control
The hysteresis control method was developed to meet the power requirements of even faster load transient response of load elements, such as the CPU and FPGA. Because it performs controls by detecting ripples in the output, this method is also referred to as a ripple control method. The method directly monitors the output voltage by means of a comparator without going through an error amp. When detecting that the output voltage has exceeded or fallen below a set threshold level, the comparator directly turns the switch on/off. The two control schemes are available: detecting a voltage below the threshold level with a fixed on-time, and detecting above the threshold with a fixed off-time.
This method offers the advantages of extremely fast transient responses due to the direct control exerted by a comparator and the elimination of the need for phase compensation. The method suffers from the problems of variable switching frequencies, large jitter, and the need for an output capacitor with a relatively large equivalent series resistor (ESR) for output ripple detection. However, innovations in these areas have advanced, and increasing ICs are incorporating this method. As an example of an improved hysteresis control IC, the ripples occurring in the output are fed back in the IC so that the ceramic capacitor with a small ESR value can be used to minimize output ripples.
・Evaluate the features and pros/cons of each control method and select the method best suited for a given design.