Project

Ultra-high-performance digital-only analog-to-digital converters through time/frequency/phase domain optimization

Code
3G0A0819
Duration
01 January 2019 → 31 December 2022
Funding
Research Foundation - Flanders (FWO)
Research disciplines
  • Engineering and technology
    • Analogue, RF and mixed signal integrated circuits
Keywords
analog-to-digital
 
Project description

Although today's electronic systems implement most of their functionality in the digital domain,
several analog operations remain necessary, e.g. amplification, filtering and above all analog-todigital
conversion. These analog circuits are preferably co-integrated with the digital circuits in
today's nanometer CMOS technology. But this gives two important problems. First, the analog
transistor performance (e.g. gain) is poor in such technologies. Secondly, the supply voltage has
shrunk to 1 V or less, which dramatically limits the signal swing. As noise sources remain unchanged,
and mismatches and interferences by nearby digital circuits are even increasing, the dynamic range
is severely limited. This makes analog circuit design increasingly challenging. As a result, when
moving to a more advanced CMOS technology, analog circuits do not become smaller in chip area,
so not cheaper.
Therefore, the traditional way of building analog circuits needs to be rethought fundamentally
towards a scalable, digital-only approach. Focus will be on the analog-to-digital converters. An
important first step were the VCO-ADCs, but their performance is still limited today. The project's
goal is to drastically boost the performance of digital-only solutions, both in bandwidth and in
accuracy, by exploring innovative architectures. The basic idea is to use signal representations that
can be represented by 2-level ("digital") waveforms and to optimally exploit the
time/frequency/phase domain.