Before Selecting A Microcontroller, Ask Yourself These Seven Questions

Jan. 23, 2007
Engineers often get more than they bargained for when they’re specifying microcontrollers (MCUs). Considering the thousands of MCU varieties available, it can be difficult to figure out which MCU is right for an application. MCUs range from products that

Engineers often get more than they bargained for when they’re specifying microcontrollers (MCUs). Considering the thousands of MCU varieties available, it can be difficult to figure out which MCU is right for an application. MCUs range from products that cost less than 20 cents for very simple home applications to $20 32-bit MCUs that control major components of industry machinery.

The natural tendency to err on the side of greater performance or more memory may, in fact, result in needless extra cost and complexity. The lines between 8-bit, 16-bit, and 32-bit products are increasingly blurred. As a result, engineers have to ask a much broader range of questions than ever before to determine the MCU with the best performance/cost qualities for their specific application, and this trend will only continue as MCUs gain new capabilities. Before making these decisions, designers should ask themselves seven key questions.

Which architecture better fits my application, CISC or RISC?

As performance demands increase, customers tend to immediately switch from the CISC to the RISC architecture. New MCUs based on the CISC architecture run at 80 MHz to achieve 80 MIPS. For applications where there is a software legacy, or in applications such as listening to music or e-learning where music or speech must be clear and high quality, these new CISC product families provide an attractive option.

What other functions can the MCU take on?

The integration of more features into the MCU is an ongoing trend. Typical features that previously were external but not integrated include DSP functionality, power on reset (POR), and low-voltage detection (LVD). Customers should ask what other features the supplier can integrate now and plans to integrate in the future to determine if that supplier can meet the customer’s long-term needs. For example, an on-board debugging feature enables engineers to use the MCU in a target system during the development phase to access all registers internally. In addition, some customers use on-board debugging to find errors in the field.

What are each MCU’s erase/rewrite rates?

Customers have commented that sometimes there are disadvantages to having more memory on an MCU. Since the MCU typically requires programming two or three times during production, erase/rewrite speed becomes important. Newer classes of MCUs with up to 512 kbytes of flash memory can be erased and rewritten in as little as two to three seconds, compared to other MCUs that can require as much as 40 seconds.

Do core and flash performance match up?

With many MCUs, flash memory is slower than core memory, which results in the MCU copying flash memory into RAM and then executing from RAM to keep up. Engineers should look for MCUs where core performance and flash performance are matched.

What is the right amount of memory?

The movement to increasing memory size and the adoption of flash memory are givens. Increasing memory size enables customers to request an ever expanding range of features, such as networking interfaces, graphic controllers, LCD controllers, multiply-accumulate (MAC) operation, and timer and I/O engines. The abilities to retain data after removing power and rapidly reprogram products have driven the move to flash. In the short term, flash has increased product price. Over the long term, these prices will start to decline.

What are the pricing dynamics for the MCU I require?

The pricing delta for 8-bit, 16-bit, and 32-bit MCUs has shrunk to the point that in many cases, only a few cents separate each.

Does the supplier fit my short-term and long-term needs?

Customers should focus on several characteristics of the supplier in addition to the quality of the specific MCU. For instance, what are the supplier’s failure rates in the field and in production? Low failure rates translate to lower long-term costs to the customer for the MCU. Suppliers should develop a broad line of products that overlaps with the customer’s organization-wide needs so the customer can purchase a wide range of MCUs, displays, and other products from a single supplier. The customer should ask about the industries and geographic areas where the supplier is strongest to ensure that products match specific customer needs and to determine if the supplier is dedicated to a given market or area for the long term. Finally, the supplier should possess production redundancy—i.e., two or more factories should produce a given MCU so the customer can continue to receive product if one factory is shut down.

The MCU market will continue to enjoy attractive growth rates, ranging from just over 5% percent for 8-bit products to more than 9% for 32-bit microcontrollers, notes Gartner Group. Trends including the opening of new markets such as China for home appliances as well as the proliferation of electronic systems in automobiles will drive volume and innovation for these processors. Customers that ask carefully considered questions about specific technologies, markets, and other needs will be best equipped to leverage the exciting developments that are taking place.

Roland Gehrmann is manager of business development in the ASSP Business Unit at Toshiba America Electronic Components Inc. (TAEC). He is responsible for all marketing for 8- to 32-bit embedded microcontrollers. Also, he has an electrical engineering degree from the University of Muenster, Germany.

Sponsored Recommendations


To join the conversation, and become an exclusive member of Electronic Design, create an account today!