Programming Mode
The compensation of current-mode buck converters is simpler than that of voltage-mode boost and flyback converters due to the single pole of the power stage. The control to inductor transfer function boost converter can be simplified as following:
Where system loop. The control to inductor transfer function flyback converter can also be simplified as following:
Fig. 33. Inductor Current with Compensation Ramp in DC-DC Converter.
2 )
Where
, which are similar to pole-zero system to boost converter, and compensation technique would focus on proportion-integral compensator applied to these two-channel system.
Specifically, only the proportion-integral (PI) compensator is utilized to enhance the low-frequency gain. The compensation zero is usually used to cancel the system pole ωps, which is located at the output node. The choice of the compensation zero will affect the performance of current-mode buck converters due to the load variations. Compared to the design of voltage-mode buck converters, the complexity of the compensation is reduced but the system compensation becomes difficult due to the characteristic of the system pole’s load dependence. Owing to the variations of load current, the PI compensation technique generates a fixed pole-zero pair to raise the low-frequency gain at the expense of transient performance.
As shown in Fig. 34(a), the compensation zero is designed to cancel the system pole at heavy loads. When load current changes from heavy to light, the system pole moves toward the origin due to its characteristic of load dependence. The phase margin worsens and the system may thus suffer from the ringing problem owing to the lesser phase margin at light loads, that is, since the compensation zero cannot be moved back to cancel the effect of the system pole at light loads, the phase margin of the system is deteriorated at light loads. The transient response shows the ringing problem and the long settling time when the load current changes from heavy to light. On the other hand, what happens when the compensation zero is used to cancel the system pole at light loads? As depicted in Fig. 34(b), if the compensation zero is designed at the low-frequency position and the system pole moves to a higher
frequency position in case of the heavy load current, then the phase margin becomes larger than the optimal value for achieving fast transient response. The compensation zero causes the phase margin to be larger than 90 degree. The system is always stable but the transient response is not good at the heavy load condition. As shown in Fig. 34(b), the transient response is slow due to the large PM value.
Therefore, an adaptive compensation zero is needed to ensure a suitable system bandwidth and phase margin. The compensation zero should be located at a low-frequency position to cancel the effect of the system pole at light loads. Likewise, the compensation zero should be moved toward a high-frequency position to cancel the effect of the system pole at heavy loads. Thus, in order to get an adaptive phase margin, an adaptive zero is used to keep a suitable phase margin value. However, the adaptive zero is only suitable for improving the system performance in steady state because the adjustment of the adaptive compensation zero depends on the values of load current and output voltage. At the beginning of the load transient response, the compensation zero is located at a low-frequency position assuming that the load current changes from light to heavy. Thus, the bandwidth of the system becomes large at the beginning of the load transient response because the system pole moves toward a higher frequency position. Once the adaptive zero is moved to a higher frequency position for the cancellation of the system pole, the bandwidth becomes smaller than that at the beginning of the load transient response. In other words, the recovery period becomes longer owing to the small bandwidth. In order to get fast transient response, the fast transient technique is proposed to accompany with the utilization of the adaptive compensation zero. To summarize the purpose of the paper, the adaptive compensation capacitance is proposed to achieve fast transient performance. In the meanwhile, adaptive resistance is utilized to enhance the performance of the converters at steady state.
During load transient period, the voltage drop depends on ESR resistance of the output capacitance and loop bandwidth of the system. The drop voltage across the ESR resistance is a material-related value. Furthermore, the system bandwidth and phase margin are important design issues in minimizing the transient voltage drop. How to maintain a high system bandwidth and suitable phase margin is therefore an urgent problem for power management IC designers. This paper proposes a fast-transient control with adaptive phase margin technique to effectively move the compensation pole-zero pair in case of load variations. At the beginning of the load transient response, the adaptive compensation capacitance (CC) is compensation zero designed at a high-frequency position to cancel the system at heavy loads causes the phase margin to worsen at light loads. (b) The compensation zero designed at a low-frequency position to cancel the system at light loads causes the phase margin to become larger than the optimum value.
decreased to move the compensation pole-zero pair to a higher frequency in order to achieve a fast transient response [47], [48]. At the end of the load transient response, the pole-zero pair is moved back to an optimal position in order to extend the system bandwidth and phase margin based on the instant load condition. Hence, the combination the adaptive capacitance CC and adaptive resistance RZ circuits for current mode DC-DC converters has good transient performance and, at the same time, ensures stability.