High-speed serial input-output (I/O) link has required advanced equalization and modulation techniques to mitigate inter-symbol interference (ISI) caused by multi-Gb/s signaling over band-limited channels. Increasing demands for transceiver power and area complexity has leveraged on-going interest in analog-to-digital converter (ADC) based link, which allows for robust equalization and flexible adaptation to advanced signaling. With diverse options in ISI control techniques, link performance analysis for complicated transceiver architectures is very important. This work presents advanced statistical modeling for ADC-based link, performance comparison of existing modulation and equalization techniques, and proposed hybrid ADC-based receiver that achieves further power saving in digital equalization. Statistical analysis precisely estimates high-speed link margins at given implementation constrains and low target bit-error-rate (BER), typically ranges from 1e-12 to 1e-15, by applying proper statistical bound of noise and distortion. The proposed statistical ADC-based link modeling utilizes bounded probability density function (PDF) of limited quantization distortion (4-6 bits) through digital feed-forward and decision feedback equalizers (FFE-DFE) to improve low target BER estimation. Based on statistical modeling, this work surveys the impact of insufficient equalization, jitter and crosstalk on modulation selection among two and four level pulse amplitude modulation (PAM-2 and PAM-4, respectively) and duobinary, and ADC resolution reduction performance by partial analog equalizer (PAE). While the information of channel loss at effective Nyquist frequency and signaling constellation loss initially guides modulation selection, the statistical analysis results show that PAM-4 best tolerates jitter and crosstalk, and duobinary requires the least equalization complexity. Meanwhile, despite robust digital equalization, high-speed ADC complexity and power consumption is still a critical bottleneck, so that PAE is necessitated to reduce ADC resolution requirement. Statistical analysis presents up to 8-bit resolution is required in 12.5Gb/s data communications at 46dB of channel loss without PAE, while 5-bit ADC is enough with 3-tap FFE PAE. For optimal ADC resolution reduction by PAE, digital equalizer complexity also increases to provide enough margin tolerating significant quantization distortion. The proposed hybrid receiver defines unreliable signal thresholds by statistical analysis and selectively takes additional digital equalization to save potentially increasing dynamic power consumption in digital. Simulation results report that the hybrid receiver saves at least 64% of digital equalization power with 3-tap FFE PAE in 12.5Gb/s data rate and up to 46dB loss channels. Finally, this work shows the use of embedded-DFE ADC in the hybrid receiver is limited by error propagation.

Inter-chip input-output (I/O) communication bandwidth demand, which rapidly scaled with integrated circuit scaling, has leveraged equalization techniques to operate reliably on band-limited channels at additional power and area complexity. High-bandwidth inter-chip optical interconnect architectures have the potential to address this increasing I/O bandwidth. Considering future tera-scale systems, power dissipation of the high-speed I/O link becomes a significant concern. This work presents a design flow for the power optimization and comparison of high-speed electrical and optical links at a given data rate and channel type in 90 nm and 45 nm CMOS technologies. The electrical I/O design framework combines statistical link analysis techniques, which are used to determine the link margins at a given bit-error rate (BER), with circuit power estimates based on normalized transistor parameters extracted with a constant current density methodology to predict the power-optimum equalization architecture, circuit style, and transmit swing at a given data rate and process node for three different channels. The transmitter output swing is scaled to operate the link at optimal power efficiency. Under consideration for optical links are a near-term architecture consisting of discrete vertical-cavity surface-emitting lasers (VCSEL) with p-i-n photodetectors (PD) and three long-term integrated photonic architectures that use waveguide metal-semiconductor-metal (MSM) photodetectors and either electro-absorption modulator (EAM), ring resonator modulator (RRM), or Mach-Zehnder modulator (MZM) sources. The normalized transistor parameters are applied to jointly optimize the transmitter and receiver circuitry to minimize total optical link power dissipation for a specified data rate and process technology at a given BER. Analysis results shows that low loss channel characteristics and minimal circuit complexity, together with scaling of transmitter output swing, allows electrical links to achieve excellent power efficiency at high data rates. While the high-loss channel is primarily limited by severe frequency dependent losses to 12 Gb/s, the critical timing path of the first tap of the decision feedback equalizer (DFE) limits the operation of low-loss channels above 20 Gb/s. Among the optical links, the VCSEL-based link is limited by its bandwidth and maximum power levels to a data rate of 24 Gb/s whereas EAM and RRM are both attractive integrated photonic technologies capable of scaling data rates past 30 Gb/s achieving excellent power efficiency in the 45 nm node and are primarily limited by coupling and device insertion losses. While MZM offers robust operation due to its wide optical bandwidth, significant improvements in power efficiency must be achieved to become applicable for high density applications.