Configurable mixed-signal ICs and asynchronous state machines can optimize embedded designs

February 07, 2017 // By Mike Noonen, John McDonald, Nathan John
Configurable ICs can make lower power, more secure and easier-to-implement alternatives to MCUs, argue executives from Silego Technologies.

Abstract – There are always external components surrounding an embedded System on a Chip. This paper describes a new category of semiconductors based on advanced circuit design and non-volatile memory, Configurable Mixed-signal ICs, (CMICs). CMIC offer a better way to implement the unavoidable standard analog, logic and discretes. In many cases, embedded microcontrollers can be replaced with much lower power and more secure CMIC Asynchronous State Machines.

Keywords: Configurable Mixed-signal IC, Asynchronous State Machine, System on a Chip, CMIC, ASM, SoC, MCU

 

I. Introduction

Systems on a Chip for embedded devices offer an astounding level of integration. Advanced processes give SoC and MCU developers plenty of transistors to work with. These devices can integrate multi-core processors, wireless connectivity, memory as well as graphics controllers.

However, even the most sophisticated and highly integrated SoC or MCU requires some external circuitry for power management, human interface or connecting to sensors. As a result, there are almost always comparators, op amps, level shifters, various logic and discrete transistors scattered across a design. These SoCs are almost never truly systems on a chip.

In some cases, the support logic needed can be swept up into a low-end FPGA. But usually this is a expensive addition to a bill of materials and is not a cost saving over discrete components. It is also an inadequate solution since an FPGA cannot address the entire external circuit, namely the analog or discrete components.

For an embedded device, this challenge will be even more pronounced as an MCU or SoC cannot address all the possible sensor, power, and connectivity options. This is further complicated by the fact that any one embedded device will be much lower- volume than an SoC for a mobile phone application. Therefore, a typical MCU or SoC vendor will not be justified in spending the large sums needed to design and fabricate a device to support all the necessary permutations and integrate the required surrounding support circuitry.

So, are designers forced to put up with sub-optimal designs with stray logic, overpriced analog and space-consuming discretes? Will the next generation of embedded devices surrender valuable space (which takes away from more valuable battery volume) and be burdened by a bloated bill of materials?

Next: CMICs

Design category: