As promised, here's a discussion I had with STMicroelectronics (ST) on its new 8-bit microcontroller, the STM8S -- the STM8S105 and STM8S207 MCUs for industrial and consumer applications. The discussion focused on how MCUs are now beginning to shape the embedded world.
Addressing this specific query, Patrice Hamard,8-bit Product Line Marketing Manager, STMicroelectronics, said that ST is reshaping the microcontroller with a solid offer on 8bit that has a strong overlap with STM32 in terms of feature and price. "Therefore, we are going to cover the need for embedded functions with only two architectures. Compared to the previous segmentations (8-, 16- and 32-bit), we are changing it to become 8- and 32-bit only," he clarified.
On the STM8S, Hamard said that the key application areas addressed by the MCU are industrial and appliances in consistent with the robustness and the reliability. He said: "The STM8S family is supporting 5V as well as 3V, thereby making it ideal for the platform evolution as well as a good offer for the consumer and mass market. The cost advantage given with the fine lithography also allows us to propose this family to key customers in PC peripherals and consumer applications."
Rich feature set an imperative in MCUs
Rich feature set is an imperative in the MCU market. How is the STM8S meeting this requirement?
According to him, the feature set is driven by the need to reduce the bill of materials (BoM). The robustness allows simple design and board layout with less filtering. The clock controller gives low noise emission figure, thereby reducing the need for shielding. The precise clock allows the suppression of the external resonator. The embedded true E²Data suppresses the need for additional E²PROM. Safe reset (no grey area) makes the reset system safer suppressing the need for external reset circuit.
The clock system, as well as the two independent watchdogs will contribute to pass safety regulations together with ST's class B libraries. All communications peripherals are available as well -- (U(S)ART, I²C, SPI, CAN, LIN), advanced 16-bit timers and timebase, fast and precise 10-bit ADC.
Finally, the 8-bit core is one of the most efficient with 20MIPS at 24MHz. Built around the 8-bit data path, the micro has 16bit registers and 32bit memory memory width.
So, how does the STM8 deliver high performance with excellent code compactness?
Hamard said that thanks to the new CISC instruction set designed in collaboration with ST's C compiler partners, the compactness has been significantly improved. The Harvard architecture with its three-stage pipeline allows to reach up to 20MIPs @ 24MHz.
ST is offering family demonstration boards and instrument cluster reference designs as well. In fact, there are currently solutions available in ST with the STM8S/128-EVAL, as well as with third parties like raisonance with the REVA KIT. Many reference designs are complete or in progress demonstrating motor control (sensorless brushless DC motors), power management, smart card protocol, capacitive sensing, etc.
Demand for low-power MCUs
According to Hamard, the trend of low power is coming from the increase of the application base on battery in consumer and personal care, combined with a strong demand for power meters (electricity, water and gas). Energy saving is important and electronics can contribute a great deal to reduce the overall energy consumption.
"The STM8S is not specifically aiming low power applications even though the features of the family are good for many low power devices. It is in our plan to introduce later this year a dedicated family to address low voltage/low power arena," he added.
Why 8-bit?
Considering that there are 8- vs. 16- vs. 32- bit MCUs, and also that affordable prices are perhaps the reason that the Asian region is witnessing a migration to 16-bit architectures. In this scenario, why 8-bit?
Hamard said: "Everything depends on what we consider to be "affordable" and who we are talking to. For large quantity and simple functions, affordability is between $0.20 cents to $0.50 cents. By construction, a 16-bit device cannot be as effective as an 8-bit product. We even believe that the microcontroller prices will decrease and address applications served with few discrete devices. The main reason is the consistency of architecture.
"The construction of the 32- and 16- are very similar, especially with the new generation of ARM-based products. The only reason to go from 8- to 16-bit is for performance improvement. We say that our 32-bit portfolio is already overlapping the 8-bit segment in performance and in price, leaving no room for the third core structure.
"Taking a closer look at our portfolio, you will realize that our 32-bit is also providing 16-bit instruction set, and our 8-bit is built with 16-bit register, 24-bit memory address bus, etc."
Saturday, March 21, 2009
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.