{"title":"27.6 Background Capacitor-Current-Sensor Calibration of DC-DC Buck Converter with DVS for Accurately Accelerating Load-Transient Response","authors":"T. Kuo, Yi-Wei Huang, Pai-Yi Wang","doi":"10.1109/ISSCC.2019.8662296","DOIUrl":null,"url":null,"abstract":"Switching buck converters with dynamic voltage scaling (DVS) for high-efficiency high-performance computing applications need to reduce the output-voltage undershoot/overshoot ($V_{\\mathrm {US}}$/$V_{\\mathrm {OS}}$) and settling time $t_{\\mathrm {S}}$ under a large and fast-changing load current ($I_{\\mathrm {load}}$). A multiphase topology with a fast load-transient response meets these requirements. The load-transient response can be accurately accelerated to reduce $V_{\\mathrm {US}}$/$V_{\\mathrm {OS}}$ and $t_{\\mathrm {S}}$ to near their ideal values by measuring the output-capacitor current $I_{\\mathrm {C}\\mathrm {o}}$ to control the inductor’s energizing and de-energizing times, since $I_{\\mathrm {C}\\mathrm {o}}$ instantly reflects the load-current transients. An integrated capacitor-current sensor (CCS) [1] can be used to sense $I_{\\mathrm {C}\\mathrm {o}}$ by emulating the output-capacitor impedance $Z_{\\mathrm {C}\\mathrm {o}}$: comprising capacitance $C_{\\mathrm{O}}$, the equivalent series resistance $R_{\\mathrm {E}\\mathrm {S}\\mathrm {R}}$, and inductance $L_{\\mathrm {E}\\mathrm {S}\\mathrm {L}}$. However, $I_{\\mathrm {C}\\mathrm {o}}$ will be inaccurately sensed if $Z_{\\mathrm {C}\\mathrm {o}}$ varies with different output voltages $V_{\\mathrm{O}}$, manufacturing variations, PCB parasitics, temperature, and aging. The state-of-the-art CCS calibration technique [1] for such $Z_{\\mathrm {C}\\mathrm {o}}$ variations is suitable for foreground operation and DVS with pre-characterized $V_{\\mathrm{O}}$ levels, since calibration starts immediately after being enabled and runs continuously until it ends. The CCS in [1] is calibrated with a low-power cost-effective comparator and successive approximation logic, with an acceptable calibration time $T_{\\mathrm{CAL}}$ for foreground operation. To broaden the range of applications, this work proposes an ADC-based CCS and a background CCS calibration (BCC) controller. The proposed CCS uses a flash ADC with a dynamic reference to shorten $T_{\\mathrm {C}\\mathrm {A}\\mathrm {L}}$. The BCC controller automatically finds a quasi-steady state (OS), namely a short period of steady-state behavior when there is no load transient or DVS event, to trigger CCS calibration, and can interrupt CCS calibration when a load transient or a DVS event occurs. Since OSs generally exist, the BCC with a short $T_{\\mathrm{CAL}}$ can increase the flexibility of scheduling both load transients and DVS events. Thus, it is suitable for DVS with numerous $V_{\\mathrm{O}}$ levels that account for in situ parameter variations.","PeriodicalId":265551,"journal":{"name":"2019 IEEE International Solid- State Circuits Conference - (ISSCC)","volume":"88 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-03-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 IEEE International Solid- State Circuits Conference - (ISSCC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ISSCC.2019.8662296","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 2
Abstract
Switching buck converters with dynamic voltage scaling (DVS) for high-efficiency high-performance computing applications need to reduce the output-voltage undershoot/overshoot ($V_{\mathrm {US}}$/$V_{\mathrm {OS}}$) and settling time $t_{\mathrm {S}}$ under a large and fast-changing load current ($I_{\mathrm {load}}$). A multiphase topology with a fast load-transient response meets these requirements. The load-transient response can be accurately accelerated to reduce $V_{\mathrm {US}}$/$V_{\mathrm {OS}}$ and $t_{\mathrm {S}}$ to near their ideal values by measuring the output-capacitor current $I_{\mathrm {C}\mathrm {o}}$ to control the inductor’s energizing and de-energizing times, since $I_{\mathrm {C}\mathrm {o}}$ instantly reflects the load-current transients. An integrated capacitor-current sensor (CCS) [1] can be used to sense $I_{\mathrm {C}\mathrm {o}}$ by emulating the output-capacitor impedance $Z_{\mathrm {C}\mathrm {o}}$: comprising capacitance $C_{\mathrm{O}}$, the equivalent series resistance $R_{\mathrm {E}\mathrm {S}\mathrm {R}}$, and inductance $L_{\mathrm {E}\mathrm {S}\mathrm {L}}$. However, $I_{\mathrm {C}\mathrm {o}}$ will be inaccurately sensed if $Z_{\mathrm {C}\mathrm {o}}$ varies with different output voltages $V_{\mathrm{O}}$, manufacturing variations, PCB parasitics, temperature, and aging. The state-of-the-art CCS calibration technique [1] for such $Z_{\mathrm {C}\mathrm {o}}$ variations is suitable for foreground operation and DVS with pre-characterized $V_{\mathrm{O}}$ levels, since calibration starts immediately after being enabled and runs continuously until it ends. The CCS in [1] is calibrated with a low-power cost-effective comparator and successive approximation logic, with an acceptable calibration time $T_{\mathrm{CAL}}$ for foreground operation. To broaden the range of applications, this work proposes an ADC-based CCS and a background CCS calibration (BCC) controller. The proposed CCS uses a flash ADC with a dynamic reference to shorten $T_{\mathrm {C}\mathrm {A}\mathrm {L}}$. The BCC controller automatically finds a quasi-steady state (OS), namely a short period of steady-state behavior when there is no load transient or DVS event, to trigger CCS calibration, and can interrupt CCS calibration when a load transient or a DVS event occurs. Since OSs generally exist, the BCC with a short $T_{\mathrm{CAL}}$ can increase the flexibility of scheduling both load transients and DVS events. Thus, it is suitable for DVS with numerous $V_{\mathrm{O}}$ levels that account for in situ parameter variations.