An Event-Driven Analog-to-Digital Converter For Time-Sparse Signal Sampling

In recent years,our living standards have been improved greatly from the increasingly applications of sensor networks.In some certain applications(such as:the weather forecast?earth's crust detection?implantable biological signal detection),the sensors are powered by batteries.Hence the power consumption is a key consideration in those sensors.The ultra-low power system,in such application,can be view as a reasonable solution to increase the working period.The analog to digital converter(ADC)is an indispensable component in sensor networks,which determines the following data transmission and signal processing.Thus,the power con-sumption and sampling method will directly influence the sensor performance.Hence it acts as a key to lower the system power.By the short time fourier transform analysis,it is very clear that the bandwidths of the signal at different time segments are different.So traditional Nyquist-based sampling method is unsuitable.Nevertheless,the event-driven ADC,sampling the signal from the time-domain viewpoint,is very suitable for the time-sparse signal sampling for the utilization of the signal's activity.In this thesis,after a deep analysis,the number of the comparator was reduced to only one.Based on one comparator structure,two different kinds of event-driven ADCs have been pro-posed.One of the proposed work,adaptive event-driven ADC,was taped out by CMOS 0.18?m Mixed-signal 1P6M technique.A single tone sine wave at 50Hz is the input when the supple voltage is 0.8V.The simulation power consumption is 50nW,chip area is 0.03mm~2,effec-tive number of bit(ENOB)is 5.8bits.The maximum frequency is 10k Hz.Within the similar type ADC,presented by the world famous journals,the best power consumption is 220nW,and the smallest effective chip area is 0.045mm~2.Additionally,another simpler event-driven ADC structure was also introduced,which can also do the right signal sampling for the time-sparse signals.