5G systems have been a proving ground for advanced RF electronic design automation (EDA) tools. These tools showcased multi-domain simulation with accurate, robust behavioral models and authentic waveforms. Teams embracing design-for-context in a trusted, high-fidelity virtual workspace were able to address challenges more efficiently with fewer hardware re-spins.

6G brings a significant multi-dimensional leap in performance requirements for components, modules and systems. Everything we think is known today about real-world mmWave effects such as complex modulation, channel propagation, signal-to-noise, energy efficiency and power delivery, starts to look different. Figure 1 illustrates potential sub-Terahertz (THz) frequency ranges with higher contiguous bandwidths targeted by 6G researchers, although specifications are still in flux.

Figure 16 G sub-THz frequencies have substantially wider bandwidth than previous wireless generations.

One immediate concern jumps out; precious little commercial hardware is available in these higher frequency ranges today. Hardware is scarce in D-Band (110 to 170 GHz), but almost none ventures into H-Band (220 to 330 GHz). Furthermore, much of the prerequisite advanced science, including semiconductor processes, test and measurement equipment, modeling and simulation technology and various facets of artificial intelligence (AI), is not entirely in place yet to cope with lofty 6G expectations.

The implications for RF EDA are immense. Research projects and proof-of-concept designs that once relied on some simulation before serious physical prototyping efforts will, out of necessity, shift significant resources into the virtual space until a broader selection of components and test and measurement gear arrives. Eventually, everything will be measured, but not before the next generation multi-domain simulation delivers crucial insights needed for predictable hardware designs meeting stringent requirements.

Several years ahead of formal 6G specification releases and perhaps as much as a decade before initial 6G network rollouts, predicting all the demands on RF EDA workflows remains a topic for debate. However, progress in mmWave EDA helps identify areas where sub-THz innovation, preceding earnest 6G system design initiatives, will prove extremely valuable. Prime examples include channel modeling, mixed-signal contexts and scalable, enterprise-class solutions to 6G challenges.

CHANNEL SOUNDING AND END-TO-END CHANNEL MODELING

Figure 2 Custom modulation driving channel sounding to determine impulse response and other metrics.

Figure 3 Channel sounding results portrayed by the 89600 VSA software in both frequency and time domains.

Wireless communication systems once enjoyed the luxury of selecting the optimum carrier frequency for propagation through the air. This was before spectrum scarcity and complex digital modulation schemes emerged. Managing channel behavior meant creating a link budget with enough margin to overcome routine disruptions.

6G transforms the channel equation. Data rate expectations for a single connection rise to hundreds of gigabytes per second, connection density grows to millions of devices per square kilometer and lower frequencies are occupied, forcing systems into the sub-THz spectrum. The environment is not ideal. Propagation losses climb as frequencies move higher in sub-THz ranges. Reflections from objects like buildings, vegetation and terrain add to atmospheric effects like rain and dense fog. When designs operate in lower signal-to-noise environments with tighter link budgets due to transmit power constraints, any interference creates the potential for error vector magnitude (EVM) degradation and a corresponding jump in bit error rates.

Dynamic channel evaluation and modeling become crucial tasks for 6G research and are not as simple as they might seem. Carrier frequencies, transmit power, the number of transmit channels in simultaneous use and the modulation scheme and waveforms influence the outcome. At this stage, 6G waveforms remain unknown, but with insight from 802.11ay, it becomes possible to synthesize a broadband sub-THz waveform with the appropriate complementary cumulative distribution function (CCDF) and other properties needed to explore options.

Deriving useful 6G virtual channel models starts with physical measurements. Channel sounding measures impulse response by sending a complex signal into the channel, capturing it after channel effects and comparing the results. Antenna configurations and reflective paths staged in the environment contribute to an understanding of the impairments. Figure 2 depicts a basic measurement setup with an anechoic chamber that provides a best-case environment without atmospheric effects.

Pre-configured routines in the PathWave Vector Signal Analysis (89600 VSA) software automate channel sounding. Figure 3 highlights the channel impulse response (bottom trace) with a time-domain view which helps assess delays, reflections and phasing.

Ensuring alignment between simulation and test instrumentation is vital for reliable results. A notable advantage of the Keysight RF EDA environment is its use of the same analysis engines, the core measurement science, from corresponding test and measurement platforms. PathWave System Design reuses waveforms from PathWave Signal Generation and integrates measurement feedback from the PathWave 89600 VSA and other analyzers. Virtual models created in PathWave System Design can represent 6G RF signal chains end-to-end, including detailed channel modeling as shown in Figure 4.

Figure 4 6G research begins on a virtual platform, allowing the exploration of modulation options against channel models.

Early 6G research reveals accurate performance simulation demands, comprehensive modeling and digitally-driven control of the antenna structures in lockstep with the RF signal chain. Decisions unfold on a millisecond scale, adapting the configuration and processing as channel behavior shifts and devices move. One challenging modeling problem is capturing more interaction between power amplifiers and antenna elements as the antenna scans in different directions with one or more beams. PathWave Advanced Design System (ADS) models, combined with measurements from a PNA-X network analyzer, will offer a detailed view of how impedance and power amplifier efficiency change. Incorporating that model into the RF system-level simulation in PathWave System Design increases fidelity.

EVALUATING RF PERFORMANCE IN MIXED-SIGNAL CONTEXTS WITH NONLINEAR BEHAVIOR

Historically, RF design has been primarily an analog discipline. However, modern communication systems feature more mixed-signal characteristics. Digital modulation, already a mainstay, takes on increased complexity in 6G with larger constellations and tighter spacing. Demands increase on 6G RF front-ends, power amplifiers, mixers, filters, switches and other components to perform predictably across a much wider bandwidth. This makes accurate simulation crucial.

RF performance can change at the flick of a digital switch in these mixed-signal chains. Techniques such as adaptive gain control, switching between alternate signal paths under certain conditions, phased arrays and beamforming, along with digital impairment compensation are necessary to pull signals from noise more effectively. RF EDA must simultaneously address these implications:

• Simulators must adhere to complex sequences of events and keep pace with rapid changes in behavior as inputs and signal chain states vary.
• Point sampling at selected frequencies misses anomalies as bandwidth increases.
• Single-domain simulation is insufficient, with anomalies appearing as domains interact.
• Model complexity must expand to accurately portray behavior, effect detail and interactions.

The days of using a few cherry-picked frequencies for detailed simulation and validation with measurements are gone. That approach only works in narrow bandwidths when interpolation between points tracks linearly without surprises. Under wideband excitation, nonlinear behavior and cross-domain interactions between power, frequency, time, temperature, load and DC bias can combine unexpectedly to disrupt mixed-signal chains at any point across their bandwidth. (See the article “Solving EM Densification at the Point of Design” pp. 52 to 64 in the July 2022 edition of Microwave Journal for more on RF cross-domain effects and simulation with authentic signals.)