Mixed Signal ASIC Design

What a Mixed-Signal ASIC Can Do for Your System

What is a Mixed Signal ASIC?

Mixed-Signal ASIC design combines analog and digital functionality to harness the best of both worlds and deliver a single integrated circuit solution. ASICs in general, provide a number of benefits to a system solution, including, product cost reduction, reduced footprint (small size), packaging and obfuscation of intellectual property, low power operation, parts reduction, and increased reliability. Mixed-Signal dual domain design broadens the scope of application and benefit that an ASIC can offer, including the increasingly popular System on Chip (SoC) architecture where a microprocessor core(s) in the digital domain is surrounded by peripherals that perform signal interface and processing in the analog domain.

Examples of Mixed Signal Design Functions:

  • Analog to Digital Converter – ADC
  • Digital to Analog converter – DAC
  • Signal Amplification with digital gain control
  • Phase Lock Loop for
    • Frequency synthesis
    • Clock recovery
    • Spread spectrum
    • Jitter and noise reduction
    • Demodulation
  • Common-mode noise removal
  • High Voltage & Galvanic Isolation
  • Cable Driving and Receiving
  • Switched capacitor filters
  • Switch Mode Power Supplies – SMPS
  • Proportional Integral Derivative controller – PID
  • Dynamic range compression
  • Impedance transformation

Three Categories Of Mixed Signal ASIC Design

Mixed Signal processing is more than just analog and digital on the same ASIC die. Mixed-signal ASIC design can be broken up into three general categories, analog domain, digital domain, and the gray capable area of mixed domain, where the lines blur between analog and digital processing to facilitate solving some difficult signal processing challenges. It is within this blurry mixed domain that many interesting things can be accomplished in a way that is more than just the sum of the parts and will be discussed below.

Along with many capabilities, mixed signal design does come with it’s challenges. There is an art to understanding how to marry the two domains such that they can exist on the same substrate (chip) without corrupting or polluting the signal integrity of either domain. The advantages are many though, offering a wide array of solutions for, low power, low cost, highly capable integrated solutions to challenging tasks.

The Analog Domain:

A majority of the signals used for monitoring and decision making originate in the analog domain, requiring at least some initial processing in that domain. Signals from photodiodes, capacitive elements, load cells, thermal sensing elements, electrochemical cells, magnetic heads, etc., produce low level and often sensitive output that require some form of signal conditioning before being useful for analysis and decision making. Signal conditioning includes impedance buffering, amplification, filtering, and level shifting to name a few. These functions are often best accomplished in the analog domain.

The Digital Domain:

Complex decisions requiring extensive processing with flexibility of processing algorithms and storage of reference data are well served by digital processing capabilities of an embedded microcontroller core and/or state machine(s). Additional digital domain capabilities include, serial communication controllers, memory controllers, sampling clocks, counters, timers, PWM controllers, calibration and trim parameter storage, mode selection, threshold settings, device ID, and many more.

The Mixed Domain:

Mixed domain technology offers some exciting capabilities to the signal processing world that are more than just digital and analog on the same die. This includes, sophisticated loop control that is a hybrid of analog feedback with digital control, advanced sampling algorithms for high common mode signal extraction, windowed event discrimination in noisy environments, and a host of other advantageous domain blending configurations.

Mixed domain signal processing benefits complex multi-loop control and is used to implement important building blocks such as, the ever useful phase lock loop (PLL), switch-mode power supply (SMPS) controller, and Proportional Integral Derivative (PID) process controller.

Mixed domain mixed signal design empowers advanced signal sampling and processing techniques. Techniques that enable ASIC designers to sample, hold, and compare signals at controlled instances in time. The samples can be made very quickly in response to a sample clock or event, allowing samples to be synchronized or windowed in such a way as to occur only when a signal is valid or to avoid interfering noise sources.

Further signal conditioning can then be applied to the compared sample output such as, gain, level shifting, filtering, and integration, all prior to an analog to digital conversion into the digital domain. One significant benefit to this approach is the ability to extract minute variations in large common mode signals without the need for a high bit depth, high speed ADC. Thus saving significant power usage of a high-speed ADC and the design complexity of incorporating a high bit depth converter.

The Value Of A Mixed-Signal ASIC:

A large number of existing integrated circuit solutions are mixed signal devices. Everything from DVD players to industrial sensors, and from cell phones to medical equipment are using mixed signal technology. Mixed signal design brings a very powerful set of tools for solving demanding technical challenges and should be considered when approaching interface and control designs demanding, low power, small size, low cost, high reliability solutions.