A 100-MS/s 8-b CMOS subranging ADC with sustained parametric performance from 3.8 V down to 2.2 V

A 100-MS/s 8-b CMOS analog-to-digital converter (ADC) designed for very low supply voltage and power dissipation is presented. This single-ended-input ADC is based on the unified two-step subranging architecture, which processes the coarse and fine decisions in identical signal paths to maximize their matching. However, to minimize power and area, the coarse-to-fine overlap correction has been aggressively reduced to only one LSB. The ADC incorporates five established design techniques to maximize performance: bottom-plate sampling, distributed sampling, autozeroing, interpolation, and interleaving. Very low voltage operation required for a general purpose ADC was obtained with four additional and new circuit techniques. These are a dual-gain first-stage amplifier, differential T-gate boosting, a supply independent delay generator, and a digital delay-locked-loop controlled output driver. For a clock rate of 100 MS/s, 7.0 (7.3) effective bits for a 50 MHz (10 MHz) input are maintained from 3.8 V down to 2.2 V. At 2.2 V, this 100-MS/s converter dissipates 75 mW plus 9 mW for the reference ladder. For a typical supply of 2.7 V, it consumes just 1 mW per MS/s over the 10-160-MS/s clock frequency range. Differential nonlinearity below 0.5 LSB is maintained from 2.7 V down to 2.2 V, and it degrades only slightly to 0.8 LSB at 3.8-V supply. The converter is implemented in a 0.35-/spl mu/m CMOS process, with double-poly capacitors and no low-threshold devices.