Tuesday, January 07, 2014: This interview discusses the best takeaways of using low power MCUs, the benefits of using modern reduced power modes, utilising power budgeting techniques, and the benefits of selecting an MCU based on target application or on development support.
Abhishek Mutha of EFY got in touch with industry experts from Microchip, Freescale, Texas Instruments and Renesas who shared their thoughts on today’s microcontrollers, how they have evolved and how they can benefit design engineers incorporating microcontrollers in their projects. Combining all their responses, we bring to you the broader perspective.
Q. What are the greatest benefits for designers who use MCUs developed specifically for low power?
There are a variety of reasons that designers are looking for lower power MCUs. The primary reason is to extend battery life. Applications such as wireless sensors consume a lot of power processing and transmitting data that they collect. The value of the sensor increases the longer it operates, so extending the battery life becomes important.
Another reason is that designers want to do more with their MCUs, but still stay within their energy budget. In the past, when dealing with a tight budget, the MCU would have to sleep more often.
Advances in MCUs have dramatically dropped their RUN currents, allowing the MCU to RUN more often, thereby allowing more features to be designed into the application software.
The low power MCU gives designer an edge to work over various applications and get a wide range of advantages like low power libraries, gate level optimization and power gating. Performance-per-watt levels have evolved in many embedded applications to enable power efficient designs which help the end user to conserve energy. In fact, considering the looming power crisis, regulatory authorities have made it mandatory for consumer appliance companies to upgrade their products and display the energy efficiency ratings enabling the consumer to make smart choices. Energy efficiency is the need of the hour and low power MCUs are the foundation for efficient devices/appliances.
Energy efficiency and power management are the key technology trends that designers are looking for. Across all electronics devices, from consumer electronics, household appliances, manufacturing equipment and metering, low power MCUs help in increasing power efficiency of that product, by harvesting and effectively utilising the power, thus making the systems more cost effective. Low power MCUs are also important for designers to help them develop environment friendly products, as low power helps in reducing radiated emission effects.
Consumers and industrial companies alike are increasingly looking towards wireless and battery-backed instruments. It is then important for each and every component in a battery operated system to consume as little power as possible, as this will translate into a smaller power source and indirectly lower component cost and PCB size. From the MCU design point of view, recent developments in advanced low power Microcontroller technology has enabled designers to attain high MCU processing performance, and yet low power consumption in standby modes to maximize the battery life in those portable battery-operated and battery backed applications.
Q. In many applications, the controller does not run continuously and may be ‘sleeping’ much of the time. How have sleep modes evolved during the years?
Rather than having two modes, i.e. RUN and SLEEP, modern MCUs have a spectrum of modes in-between that offer varying degrees of functionality.
On the PIC® MCUs with eXtreme Low Power, the modes are named Doze, Idle, LVSleep, Sleep, and Deep Sleep. Modes like Doze allow the peripherals to be clocked at a higher rate to meet the needs of communications or timer sequences, but allow the core to run at much lower speeds, reducing the current consumption by up to 75%. Modes such as Idle turn the core off completely. Modes like LVSleep, Sleep and Deep Sleep turn off all peripherals but preserve select features like LCD, ADC, Timers and Interrupts. So a lot can be accomplished even in these modes which save 97-99.9% of the current than RUN mode.
Real-time low-energy environments pose special challenges for the designer and programmer because the same sleep states that reduce an MCU’s energy consumption often also reduce its ability to quickly respond to an event. Low-power modes typically range from a light sleep or standby mode, through deep-sleep, generally, the deeper the sleep, the less power is consumed by the MCU.
A number of strategies are used in projects to determine how these variables can be set to take best advantage of an MCU’s sleep modes.
One is to identify the routine low-level functions required by the application and identify MCU architectures that can handle as many of them as possible using on-chip hardware that does not require CPU intervention. Other methods are clock gating, gate level optimization, voltage scaling and digital architecture techniques.
Sleep modes are available on all the MCUs now days, but it depends on the architecture of the MCU and the overall design implementation (both hardware & software). Basically, if the system is designed intelligently and if the MCU Architecture supports this, power consumption can be reduced significantly. For example most MCUs have DMA now and if a data transfer function has to be done, the core can actually be kept in “OFF” Mode which saves a substantial power. Also, the time which the internal PLL or FLL takes to stabilise from low power to active mode also determines the average power consumption.
There are standard techniques of reducing power by optimising the voltage levels and clock frequency. Additionally, attention needs to be paid on the IO lines leakage (including communication interfaces) which can help in substantial power savings.
“Sleep” mode or “standby” mode has certainly evolved over the years. In Renesas’s RL78 series, we certainly showcased such evolution. In this product, we implemented 3 types of ‘sleep’ mode. They are –
1. HALT mode
2. STOP mode
3. SNOOZE mode.
SNOOZE mode is especially interesting because it allows some peripherals like ADC and UART operation while in standby modes. Besides taking advantage of the various MCU “sleep” modes available, designers can moderate the CPU performance by switching off or scaling back the clock speed whenever possible.
They can also further reduce the MCU power consumption by turning off any clocks and peripheral blocks that are not being used.
Q. Since a microcontroller can spend substantial amounts of time inactive, many designers use a ‘power budget’ to determine the average power consumption and to calculate the battery requirements. Your thoughts on this?
Microchip’s free Battery Life Estimator provides MCU models to generate a report of the average current, the life of a battery in that configuration, and if the battery will support the modelled application. This allows trying different approaches to save power like whether it is better to run a portion of their code at full speed, to finish faster, or if running slower for longer gives lowest power consumption.
(Note that this is only a model. There are a lot of other variables involved, so bench trials are always important.)
Yes, designers are using power budget these days to determine the power consumption and also to improve and calculate the battery performance and its further requirements. This is a good practice as one can check the performance of the battery and exactly know where it lags (if it does). This ensures good quality battery and also a longer life cycle.
Power budget calculation is never an easy exercise. There are many aspects that need to be considered while working out the battery life estimate. Important parameters include:
– Battery self leakage
– Duty cycle calculation (active & sleep modes)
– Time duration during switching of different modes
– Optimised voltage & frequency levels
– Peripheral circuitry – power calculations (Iq Currents of active devices)
Power budgeting technique is certainly one of the method used to estimate the average power consumption. But one should not take this as accurate and make critical recommendation of a system’s operating hours based on particular battery type. Such estimation should take into consideration of other components on-board too and designers should not make calculation from MCU aspects only.
Q. What are the most exciting features that have come up to delight engineers, be it in the software for programming the MCU or the embedded peripherals within the MCU?
Low power MCUs are now available with 16-bit Sigma-Delta ADCs built in. For communications we are seeing USB, CAN, and I2S™ being adopted in many ways. And finally, we are seeing small bits of programmable logic being added. This is great in that you can now stitch together peripherals in the manner you want within the MCU and have it operate in sleep mode at tiny current levels.
In terms of peripherals, MCU now offer high resolution ADC upto 24 Bit Sigma delta ADC, high res. DAC, USB, Ethernet, high speed PWM’s, RTC, LCD controllers etc. In terms of programming all of them offer IDE and good compilers, we also offer Real time OS free with our 32 bit MCU’s.
From a functionality perspective, we see increased analog peripherals optimised for certain end equipment or applications, being adapted by most suppliers making it a cost effective implementation (Op-Amps, 16-24bit ADCs, DACs) Also, connectivity is gaining ground across embedded applications and peripherals like Ethernet, CAN, USB and RF are increasing in most of their offerings.
Today, functions that are ‘Smart’ and ‘Connected’ are the much sort after functionalities in MCUs. For example, in Renesas RL78 Smart Analog series, analog front-end are integrated into the MCU. It is a new technology that allows interfacing to hundreds of sensor types with a single, configurable analog front end (AFE) platform – resulting in more intelligent sensors with a smaller footprint, while significantly reducing development time.
Q. “Internet of Things” is bringing about a significant wave of change in electronic devices. Are there MCUs that allow designers to easily build and link projects in the IoT space – an IoT application focused selection of MCUs perhaps?
With the advent of low power radios that run in unison with the low power MCU and as technologies like Bluetooth LE are deployed, these “things” web connected intelligence can be enabled anywhere at very low cost. For example, Microchip has worked with companies such as Powercast™ to develop battery-less sensors powered by RF energy that can be deployed anywhere within a 25 meter radius and send data to a collector that can relay to the web forever.
Internet of things is the next big revolution, giving us power to actually link anything and everything to the internet. Microcontrollers these days do allow that flexibility to the designers. Freescale’s Kinetis 32-bit microcontroller family is a present example of how microcontrollers allow designers to build and link projects in the Internet of Things.
Yes, there are many manufacturers supplying single chip embedded controller for Internet of Things, where microcontroller, RF and software (IPV4, IPV6 compliant) are offered as a complete solution. More recently, TI has launched – CC2538F512 which is a single chip implementation for the Internet of Things space.
It certainly allows designers to make their products connected today. For example, microcontroller today has integrated front-end radio frequency circuit for communications. Along with appropriate software stacks, the microcontroller ecosystem allows protocols like Bluetooth or Zigbee to be implemented in a range of applications. Among the applications are Smart Grid or Smart Home for example.
Q. Do you think it is important for designers to consider the availability of software development tools, as a major factor, for the selection of an MCU?
Applications are getting more and more sophisticated, software tools are a key consideration, if not a primary consideration, when selecting an MCU. A recent survey concluded that over 66% of design teams now consist of software engineers. This is reflective of where solutions are needed most. To get applications to market on time and within budget, time must be saved on the software development side.
Software and hardware goes hand-in-hand. Gone are the days when the companies only use to look at the hardware aspect and then customer use to source software on it. Today, companies like us are providing complete solutions to our customers which make it easier for the design engineer.
Yes, it is a key decision factor, as developer do consider a certain architecture on the available software tools & libraries built on that MCU platform. Time to market (development) is a very critical need across the industry and software availability plays a very critical role to address this.
Yes, it is one major factor that designer should take into consideration when selecting MCU. The design should not be compromised by lack of development tools or the lack of the tool’s features and functions. And the designer themselves, should not be loaded with the burden of tools related problems. At Renesas, we always assure our customers of quality development tools and making sure that they are easily available to them.
Electronics For You SOUTH ASIA’S MOST POPULAR ELECTRONICS MAGAZINE