16 or 32? Why 16-Bit is Sometimes the Right Choice
Matching application to MCUs is the key
Many people will tell you that the 16-bit microcontroller (MCU) is obsolete and that you should take the leap from 8- to 32-bit, missing out the step in the middle. These people argue that nobody needs 16-bit devices. In a way this could be true, but only in the same way that we don’t need anything other that a Smart car or a Mercedes S Class to travel from A to B. But neither of them will be the optimum choice for the vast majority of journeys.
For everyday travel a car must be fit-for-purpose with the performance, features and functions that we need, or desire. This must be at a price that’s affordable and, ideally, the lowest possible energy consumption that’s consistent with the aforementioned performance and features. It is the same with microcontrollers. The continuing popularity of the 8-bit micro (of which an early death has been regularly predicted for the last 15 years or so) demonstrates that basic functionality at modest price always has a place. It’s true that there are a growing number of applications that need 32-bit devices for top-of-the-range performance but, just as in the car analogy, there is great demand in the middle ground. Here 16-bit devices, and platforms based on these, offer the right balance of CPU performance, memory performance, peripherals, pin-count and usability to meet the majority of applications.
All building blocks within the MCU, including the core, bus width and Flash memory performance need to be in balance to achieve a cost-effective outcome - and in such a highly competitive marketplace where every penny counts, customers need an optimised design for their application. If any of these items are over-specified for a given application, it increases the cost. With this in mind, 8-bit devices will always be the most cost-effective for 8-bit applications, 16-bit devices for 16-bit applications and 32-bit devices for 32-bit applications. Renesas' MCU design philosophy aims to ensure that you can always achieve the best fit across the performance range.
Figure 1: A Platform Approach Demands that Devices are Core-, Code-, Peripheral- and Tool-Compatible and, Where Possible, Even Pin-Compatible
It is important to match MCU performance to the requirement of the application for technical reasons too. Consider for a moment power consumption and ElectroMagnetic Interference (EMI). Power consumption is related to core complexity and clock frequency. Where you have access to a wide choice of core sizes and performance, you can ensure that the lowest power device can be selected for any given application, something that’s particularly important in prolonging equipment battery life. EMI performance is much less of a problem with smaller and lower speed cores, so by selecting the best fitting 8- to 32-bit core for the application, you achieve the optimum EMI performance and you can meet CE/FCC compliance more easily than if an unnecessarily high performance core were used.
Many areas of industry are moving to a platform-based approach, where one base system design is reused in many different end equipment models. Cars and mobile phones are two examples. In these applications, cost is still a critical issue. With respect to MCUs, Renesas addresses this need by providing a line-up of devices for 8- to 32-bit applications that are core-compatible, code-compatible, peripheral-compatible, tool- and even pin-compatible. This means that designers can base a platform on one MCU but can substitute an alternative, compatible part on to the PCB to give either higher or lower performance as required for the specific equipment model. This ensures that you can achieve a cost-optimised result every time. Without a 16-bit option, it’s impossible to optimise the range of applications that we see every day in the market.
Renesas MCUs are all designed for development using 'C' programming language, with resulting highly efficient code and optimised code density. It also ensures fast and easy re-compiling to any device in the product range delivering the cost benefits of the right device for the right application. In addition, and often more important, the MCU cores have been designed for embedded-C applications, which makes them extremely efficient in embedded applications. For example, they provide extremely fast interrupt response times and powerful bit-manipulation instructions, which can often be a major limitation with other architectures in real world applications.
In most applications today, cost is critical and every cent or penny counts. Higher device integration is one important way to reduce overall system cost. By integrating key functions into an MCU line-up - for example, power-on reset, low-voltage detect, oscillators and Flash memory - external component count, PCB space, cost and manufacturing costs are all reduced. In addition, Renesas MCUs run from a single voltage supply rail, so system power supply design is simpler and lower cost than for devices that require more complex power arrangements. Renesas designs all its MCUs with system cost in mind to provide the optimum product at all different price/performance points. Furthermore, choosing a scalable CPU architecture makes it easy to integrate peripherals in software and still ensure optimised CPU performance for the application.
Flash memory is becoming a de-facto requirement for today's MCUs. Renesas has a range of Flash architectures, from industry-leading high-performance and high-reliability Flash for demanding applications, ranging to lower cost and performance-optimised Flash for less intensive applications. This range of technologies further ensures that the best-fit device can be selected for all price/performance points. Renesas Flash also has zero wait-state operation. This is a critical benefit for many applications. It means that the Flash can work in harmony with the speed of the CPU core giving maximum system performance. Unlike other architectures, CPU performance will not degrade because the CPU needs to wait for the Flash memory to respond.
As the above demonstrates, in today's highly competitive environment it is important to select the right device for the right application. By ensuring these devices are compatible across the performance range, you get many other benefits. So there will always be a strong market for a complete range of architectures from 8- to 32-bit. Many independent market surveys confirm this and forecast strong growth in all areas over the coming five years.
Upgrading from 8-bit to 32-bit, just because 32-bit is available, can be illogical, expensive and potentially a recipe for disaster.
About the author:
Tim Burgess is the European Marketing Manager for Consumer, Industrial and Distribution business at Renesas Technology Europe. He has been with Renesas for six years, including three at Hitachi Europe before the merger with Mitsubishi Electric. Prior to this, he spent 10 years at Toshiba Semiconductor in a variety of marketing, sales and applications engineering roles. He holds a BSc in Communication Engineering from Plymouth University, UK.
Arrow Electronics, Inc is a global provider of products, services and solutions to industrial and commercial users of electronic components and enterprise computing solutions.