Search

For discontinuous conduction mode (DCM) dc–dc converters with digital sensorless current mode (DSCM) control, an observed current error occurs, owing to a low-accuracy current observer. Moreover, a reference current error occurs due to a low-accuracy current controller or a finite dc gain of current loop. Conventionally, the observed current is compensated to increase current regulation accuracy, whereas the reference current error is neglected. However, for charge balance principle (CBP)-based DSCM (CBP-DSCM) control, this paper proves that the reference current should be compensated in a same quantity to that of observed current. Otherwise, single or unequal compensation leads to output voltage steady-state error. For this reason, a dual current error compensation strategy for CBP-DSCM control is proposed. It compensates the errors in a same quantity through a current error observer, which considers parasitics and calculates the current errors without approximation. To verify the proposed strategy, small-signal models with parasitics for both the converter and the controller are constructed by differential functions of key variables. Furthermore, converter stability is analyzed at typical operation condition, while the stability at various operation conditions is verified through robustness analysis. Finally, simulations and experimental results verify the analysis and the improved transient response of the converter. [ABSTRACT FROM AUTHOR]

For a sensorless predictive current controlled boost dc–dc converter, its small-signal model that contains a number of parasitic parameters, is derived in this paper. This model indicates that the type of system becomes type 0 even with the correction of voltage loop proportional–integral controller, leading to the existence of output voltage steady-state error. Then a self-correction differential current observer (SDCO) is proposed to eliminate this steady-state error and gain high transient response speed. The self-correction part of the SDCO makes the system become type 1 to achieve no steady-state error for output voltage, whereas the differential part can guarantee that the intermediate calculation results do not overflow. By carrying out a series of simulation verifications, further investigation proves that the proposed algorithm has good robustness. Finally, the effectiveness of the proposed algorithm is verified by experimental results. [ABSTRACT FROM PUBLISHER]