What is AMS Verification:
Analog Mixed Signal (AMS) refers to simulations that utilize both analog and digital solvers. In the context of integrated circuit (IC) design, AMS verification involves testing and validating analog and mixed-signal ICs as per the design requirements. AMS designs combine both analog components (e.g., amplifiers, filters, oscillators) and digital components (e.g., logic gates, microcontrollers, digital signal processors), making their verification more complex compared to purely digital systems.
AMS Verification Methodologies:
Simulation-based Verification:
Simulation is a widely used method for verifying AMS systems, including:
- Analog Simulation (Spice-based): Tools like HSPICE, Spectre, and Cadence are used to simulate the analog portions of the circuit.
- Digital Simulation (HDL-based): Tools like ModelSim and VCS are used to simulate the digital portions of the circuit.
- Mixed-Mode Simulation: Simulators such as Mentor Graphics’ Eldo, Cadence Virtuoso AMS Designer, and Synopsys’ FineSim allow for both analog and digital systems to be simulated together.
Formal Verification:
Formal verification uses mathematical methods to prove the correctness of system designs. In AMS systems, it ensures predictable behaviour and meets specifications, particularly at the interfaces between analog and digital domains. Tools like Cadence or Synopsys help verify the consistency and correctness of logic within the digital domain and ensure the accuracy of signal conversion algorithms in ADC/DAC systems.
Verification Steps Involved in AMS:
- Design the System: Model the analog components (e.g., amplifiers, ADC) and digital components (e.g., microcontrollers).
- Run Mixed-Mode Simulations: Use tools like Cadence Spectre for analog portions and a digital simulator (e.g., ModelSim) for digital components.
- Check Interface: Ensure that the ADC output is correctly mapped to the digital domain and that the DAC accurately reproduces the analog signal.
- Perform Timing Analysis: Check for timing mismatches between analog and digital components that could cause signal degradation or synchronization loss.
- Validate Signal Integrity: Use signal analysis tools to verify that converted signals (analog to digital and vice versa) maintain their integrity and accuracy.
Challenges in AMS Verification:
- Complexity: Interaction between analog and digital components increases system complexity.
- Non-linearity: Analog circuits often exhibit non-linear behaviour that is difficult to model and verify.
- Timing Issues: Synchronizing analog and digital domains can lead to issues like jitter, skew, and delay.
- Signal Integrity: Ensuring the signal is not degraded during transitions between analog and digital stages.
- Coverage: Achieving complete verification coverage across both domains is challenging due to the differing nature of analog and digital simulations.
Modelling Concepts:
Who Creates the Model: Digital Engineer or Analog Engineer?
Analog designers and digital verification engineers collaborate to create models. Digital modelling/verification experts, supported by analog designers, create models. Analog designers trained in Real Number Modelling work with verification engineers to meet verification goals.
Analog Modelling:
Analog models (Verilog-A/Verilog-AMS) provide a mathematical description of analog blocks in the continuous time domain, detailing terminals and external parameters. The analog solver resolves the value of all node voltages and branch currents.
Behavioral Modelling:
Behavioral modelling relates output signals to input signals, representing analog behaviour in a system. This approach is used to derive equations for block behaviour, enabling fast simulations typically for design purposes.
Structural Modelling:
Structural modelling describes the structure, architecture, and block parameters, modelling sub-blocks for more realistic simulations. This approach is commonly used for verification.
Real Number Modelling (RNM):
RNM uses concepts from both continuous and discrete domains. Values are floating-point (real) numbers, and time is discrete, with real signals changing values based on discrete events.
Modelling Languages Used for AMS Verification:
Verilog-AMS:
Verilog-AMS and VHDL-AMS are extensions of standard HDLs, designed to model and simulate mixed-signal systems. These languages describe the behaviour of mixed-signal circuits, including interactions between analog and digital parts, supporting both design and verification within the same framework.
Wreal Modelling:
Wreal, used in System Verilog, stands for Weighted Real, representing analog signals in digital simulators for mixed-signal verification. The wreal data type in Verilog-AMS provides continuous real values upon discrete event changes, identifying high impedance/unknown states for real values. Wreal models bridge the gap between digital and AMS systems, representing real-valued signals in digital domains.
Real Number Modelling (RNM):
RNM represents real (continuous) values in simulation environments, useful for simulating analog behaviour without traditional analog simulation tools. RNM describes analog blocks as signal-flow models, simulating them in digital solvers at near-digital speeds, using floating-point real numbers to represent voltage levels.
Conclusion
Analog Mixed-Signal (AMS) verification is a crucial process in the design and development of integrated circuits that combine both analog and digital components. The complexity of AMS verification stems from the interaction between continuous and discrete domains, necessitating advanced methodologies and tools. Simulation-based and formal verification techniques play a vital role in ensuring the correctness and integrity of AMS designs.
Modelling approaches, including analog, Behavioral , structural, and real number modelling, provide a robust framework for accurately representing and verifying mixed-signal systems. Collaboration between analog and digital engineers is essential to create comprehensive models that address the unique challenges of AMS verification.
As technology continues to evolve, the methodologies and tools for AMS verification will also advance, driving innovation and improving the reliability of mixed-signal integrated circuits. Understanding and overcoming the inherent challenges of AMS verification are key to achieving high-quality, performance-driven IC designs.
At VeriFast, we have skilled resources who can work on complex AMS verification. Do contact us if you need extended hand for AMS verification work.
