It’s been an interesting decade and I remember clearly that it was sometime in 2007 when we actually started working on head-mounted glasses project for one of our customers. The customer had an interesting idea of showing whatever you were watching on your laptop screen into the glasses and that was to be controlled and operated by voice commands. On your head, you had a hands-free computer that was controlled by your voice commands and your “hands were free” to do other tasks. If you look back nothing much has changed from that basic concept in 2007 but in reality, everything has changed.

Today we are at an inflection point for the AR, VR, MR and whatever other ‘R’ you want to call it as. What exactly has changed in last 10 years? When we started working on the glasses in those good old days, the entire idea was to realize hardware that could fit on your head, function like a computer, operate by voice command, and have Windows on it. Yes, the very first glasses that we realized had Windows Embedded OS! Android was still not in the vogue. It was a great achievement for us to show something like this at the 2009 and 2010 Consumer Electronics Show (CES) where it attracted huge crowds when it was shown for the first time. It was a standalone piece of hardware showing some basic stuff like photos, videos and such.

One thing that definitely helped the world to take notice of this technology was Google Glass. Google Glass 1.0 may not have been a critical commercial success for Google but it definitely did a great service to the smart-glasses world by showcasing the technology to the common man in a way he could understand. That was one of the biggest contributions of the Google Glass.

Coming back to today, what has changed? I think for any technology or product to succeed many things have to come together including:

1. A good hardware platform
2. Software
3. Value for money
4. And most important, real-world use cases: It should make a difference in users’ lives. It could be for business or in the day-to-day life of the common man, addressing some critical issues. At the end of the day for a product and service to succeed, it should make a significant difference in quality of life and the problem it is trying to solve.

So what do we have today that wasn’t there ten years back? Let’s take a look.

1. The hardware has matured. Various ARM™-based platforms are giving PC-grade performance
2. The software has been hardened and field tested. We actually have everything that is there in PC on these glasses too.
3. And there’s more: The apps! Today there is a community out there writing apps for the AR, VR, and MR world.

But that’s not it. We actually have the real-world use cases that can impact or make a difference to the lives of people.

The advanced research in the areas of AI, machine learning and analytics has greatly contributed to the way these products are evolving; the seamless connectivity of these glasses to the cloud is making a huge difference in the way these devices and applications can be used in the real world to solve real problems.

So what is the real-world problem that these glasses are going to solve? The possibilities are countless. I can keep listing things like factories, technical support, advanced engineering support, supply chain management, aviation, medical and even in solving a great issue like blindness by actually giving an eye to a visually challenged person. Imagine what a great improvement it would make to visually challenged people if they can start leading a normal life without any support from anybody. This is not a fantasy and we will be seeing large scale deployment of these glasses in eradicating blindness as we know today.

So how is this really going to solve the day to day issues of the various things I have mentioned above? The scope is enormous but I want to look at couple of areas where it can make a real difference.

In today’s age of e-commerce, the supply chain is probably one of the biggest challenges. For the first time, Internet shopping exceeded in-store shopping during this Thanksgiving season in the US. E-commerce in general and Amazon in particular have changed the way we shop. The inflection point has already occurred as far as online shopping is concerned. So what’s the next frontier to capture? Second-day delivery is passé. I want to touch and feel the product on the same day that I buy. Is same day delivery possible?

Have you ever imagined the logistical nightmare that moving these huge inventories cause and the amount of manual labor that’s required to accomplish this? Giants like DHL and FedEx have already started pilot projects to assess the use of AR across the supply chain ranging from warehouse planning to transportation optimization and last-mile delivery.

Imagine huge warehouses measuring hundreds of thousands of square feet and you can only guess at the time you can save if you can direct the worker to the package directly instead of looking around all over the place. It has been demonstrated that visual picking using smart glasses can improve efficiency by as much as 25 percent over regular hand picking. In regular hand picking, workers search and pick packages in the warehouse using a paper list and a handheld scanner. But in vision picking, they wear smart glasses that provide them with visual instructions via AR on what to pick and where to place the goods.

Now look at the next stage, the RFID tag is prevalent and costs no more than a few cents. Having an RFID tag on each package accompanied by a RFID reader and combined with visual instruction on glasses provides a winning solution.

Advanced remote technical support is another area that has immense potential with a return on investment that can be realized in matter of days and not months or years. Specialized industries such as oil and gas and aviation face an aging workforce. It is estimated that the aviation industry is going to lose a large number of its experts due to retirement. These experts cannot travel to remote locations to troubleshoot and fix problems. Here is where the glasses can come in handy. If there is a breakdown of equipment or a technical glitch, a specialist in a remote location can assess the situation via AR devices worn by workers present at the site and help them resolve the issue using AR features such as voice instructions and images. This equips regular employees in the field to undertake assembly and repair tasks that would otherwise require specific training effort and time.

These are just a few examples. The possibilities offered by visual reality are countless and we are just seeing the tip of the iceberg!

*Published in embedded.com

Industrial automation systems typically require data acquisition and processing in real-time. These systems could range from a simple lighting control system to a large distributed control system comprising of sensors/actuators, industrial gateways, monitoring and processing units. Designing such an embedded system puts forth a lot of challenges for design engineers in terms of striking a balance between low-end micro-controllers to performance oriented processors. This can be addressed in an optimal way if the design is done keeping the complete system and its end goals into consideration.

I’ve put forth here a high level approach to overcome these challenges using various system resources to enhance the performance.

The System

We were designing an industrial data acquisition system and had constraints in terms of the total product cost. We opted to go with a TI-Sitara processor with ADC and DAC connected over the SPI bus. The system also has the Digital IO expander connected over I2C based IO expander.  The system runs on Linux Operating system (with RT patch) with standard BSP & Drivers for all on-chip peripherals.

The key aspects of the system are: –

• System running Modbus client responding to the MODBUS master for read/write of system configuration and real-time data
• System running Web-Server to allow the control/configuration of the system and monitoring of real-time parameters
• Capable of handling up to 16 Analog input & 16 digital input.
• ADCs capable of capturing the input signals with the sampling rate of 1 KHz for each channel.
• Digital input being monitored and state changes being responded within 10 milliseconds.

System challenges

The single largest challenge determined during the implementation and testing is associated with the capture of the analog input. To be able to capture the analog input at 1KHz for all channels, the ADC has to be configured for higher sampling rate of 16KHz. Since the ADC does not provide the buffering, all the captured data will have to be read immediately to avoid it being overwritten with new data corresponding to different channel.

The ADC provides an interrupt after completing the conversion for each channel. The standard driver implementation would register the interrupt and initiate the read of the SPI registers. Incidentally, for reading the SPI registers, there is a need to perform SPI register write (so as to generate the required clock). Due to interrupt latencies of Linux (even with the RT patch incorporated), the ADC capture is being missed.

Below is the representative summary of the flow of control and data.

1. ADC completes the conversion and asserts the End of Conversion (EoC) signal
2. The Sitara registers this EoC line as interrupt and invokes the Linux GPIO Interrupt handler
3. The kernel space interrupt handler determines the source to be ADC and then accordingly invokes the ADC drivers
4. The ADC driver in kernel space issue an SPI bus read and thus reads the conversion value
5. The value is then given to the user’s space application for further processing.

Owing to the large interrupt latencies, context switching time in Linux and the interaction between the kernel space and user’s space components, there was an overall delay. This limits the overall sampling rate.

Systems approach

Issues like the one stated above are likely to occur while various disciplines are working as islands; i.e., the hardware, BSP & applications are working disconnected with each other having fixed interface plans. Instead, if the design is done keeping the complete system and its end goals into consideration, such problems can be solved in much optimal way. This is Illustrated below.

The Sitara is a complex SoC having multiple features including DMA engines. These DMA provide multiple features including event based triggering and chaining. These features have been exploited to solve the above problem. Below is the outline of control and data flow.

1. The EoC signal from the ADC is configured to trigger the DMA for issuing the SPI Write (so as to generate the clock). The clock also causes the ADC to transmit the converted data into the SPI lines all the way into the internal receive shift registers of SPI controller in Sitara
2. Being full, the internal receive shift register triggers another DMA to read the data from shift register to the memory buffer
3. Steps 1 & 2 are repeated for 16 channels. After 16 channels data is transferred to system memory by DMA, the DMA generates transfer completion interrupt
4. The interrupt handler transfers the data corresponding to 16 channels and reconfigures the DMA for next transfer waiting for EoC
5. The hardware timer is separately configured to generate the required Start of Conversion signal to ensure the sampling interval is deterministic. The ADC is configured to carry out all the channel conversion and stop (waiting for subsequent Start-of-Conversion signal).

This approach yielded the following benefits.

• The system load (CPU Load) is reduced as the Sitara is no longer executing the interrupt handler for every channel conversion completion
• The reliability of the sampling rate is hardware timer controlled and thus immune to variations on the CPU load
• Able to achieve greater sampling rates reliably.

Thus, as showcased above, a thorough understanding of the SoC and its features and seamless integration between the hardware, BSP and application are key to a successful optimization of System performance optimization for a variety of embedded devices.

*Published in EE Times India

Over the past decade the embedded world has undergone tremendous change. With the advent of mobile phones, smart lifestyle gadget like wearables, health and wellness devices, on the consumer front, users are demanding smaller and thinner gadgets.

This has lead in turn to designs that increasingly need efficient memory architectures such as high memory capacity and high performance in small area and multiple bus issues that call for high scalability.  The designs demand compact and more densely populated electronic assemblies.

Package on package (PoP) is one of the techniques to address the demand for compact assemblies. PoP is a stacked packaging method to have two ball grid array (BGA) packages mounted one above the other with a standard interface to route signals between them. The most widely used integration components for stacking are the processor and memory. The combination of RAM & Flash memory in single chip BGA solution is also available, which allows a much higher component density on smaller form factor PCB design through the PoP assembly.

Stacking memory is one way to achieve the dual goals for enhanced functionality and greater packaging density of a product. It is fast becoming a promising solution, offering high integration that leads to product miniaturization.

Mobile applications can benefit from the combination of this stacked package, offering small footprint and minimal PCB space. Other portable electronic products such as

•      mobile phones (baseband or applications processor plus combo memory),

•       digital cameras (image processor plus memory),

•       PDAs, portable media players (audio/graphics processor plus memory), gaming and others also benefit from this design approach.

Benefits of PoP Technology

Using PoP technology in a design offers many advantages. The most obvious is the reduction in the PCB size or the small footprint of the PCB. Using PoP technology also ensures a reduction in the no. of layers of the PCB as the connection lines between the processor and the memory are minimized. This also improves signal integrity on the board by minimizing trace length between different interoperating parts, such as controller and memory. The direct interconnections between the circuit yields reduced propagation delay, noise and cross-talk.  Using PoP technology also makes for easy memory scalability on the hardware. This is because most of the memory modules for PoP design come in multi-chip packages (eg: Flash + DDR). Hence, both the Flash and DDR memory can be upgraded by replacing the single PoP memory package. And finally, there is the reduction in BoM cost achieve as a result of elimination of termination discretes on the PCB.

The ARM advantage

PoP technology is most popularly being used with ARM Processor Applications. Texas Instruments was one of the first semiconductor companies that adopted this technology. This is now being followed by other silicon vendors like Freescale etc. ARM chipsets are known for low power and are popular for small footprint, portable applications. The low power ensures less thermal radiation to the memory; when it is placed over the processor in the PoP technology. In comparison, Intel SoCs and other DSP are high power and have high thermal radiation. Therefore, there is a possibility of the memory to stall during operation, if the PoP technology concept is used for these chipsets.

An important point to note for product developers is that assembly of PoP PCBs requires special skillset and the need to follow a defined assembly process. The PCB manufacturer should follow the required methodology to ensure minimal yield issues for a successful implementation to take advantage of the benefits of such a design.

*Published in EE Times India

There are a few people I consider visionaries of their times. In the 15th century, we had Leonardo da Vinci. He was centuries ahead of his times and unmatched in his abilities, talent and knowledge of subjects as varied as science, arts, materials, structures, aeronautics, civil engineering, mechanical engineering, medicine etc. A true multi-faceted genius.

In our times, we have had Steve jobs. His mix of technology & business acumen and knowing what people want has made Apple what it is today. Alas, we lost Steve….. I guess many will partially blame him for some wrong choices he made treating his illness. All said & done, he was a genius of our times. He dared to dream and dream crazy & dream big…… or as my good friend, Verne Harnish, would say he had a BHAG (Big Hairy Audacious Goal).

Moving ahead, I think we now have a living genius among us today.. ELON MUSK. He has achieved what people considered the impossible. He has a vision for the future that bridges the gap between science fiction & reality. He is changing the dynamics of technology and business. All he needs to do is dream and investors will write him a Billion Dollar check to experiment with his dream….. till he goes to sleep and dreams again.

Since he sold PayPal to eBay for $1.5B, Elon has forged ahead to bigger ideas. He built the All Electric Tesla car. No one imagined that such a car can roll out from Silicon valley. I have driven the Tesla and I consider it the best & most advanced car in the world. From cars to Rockets seems like a logical step for Elon. But I think the big disruption that he will cause in our times is in the area of Solar energy & battery technology. He will find ways that will disrupt systems that we have got accustomed to over the last 100 years. I am sure we will see many new ventures from him stable in our lifetime. Tesla Energy has already gathered a big buzz. I just can’t wait to get  a Tesla Power Wall.

Over the last 10 years, Steve Jobs enticed everyone in our homes to have 2-3 of his products which averaged a price of about $400. In the next 10, we will be buying products from Elon’s companies. The only difference is that the product prices will be anything from $3500 to $100,000. The volumes will still be in high millions.

So get ready to be mesmerized!

Consumer electronics product design is a dynamic field that blends creativity, functionality, and user-centric innovation. Consumer electronics product design aims to create devices that seamlessly integrate into users’ lives while delivering intuitive and enjoyable experiences. Design and Development of Consumer Electronics Product Designs has come a long way from traditional product development. Earlier, product development involved a semiconductor company building a reference design around a brought out silicon/processor and promoting it in the targeted market segment. Some examples include semiconductor companies like Texas Instruments and Analog Devices, who are building reference designs around their flagship processors for a variety of Lifestyle Electronics and Consumer Electronics Product Development like cell phones and automotive infotainment applications.

The Traditional Method

Reference Designs or evaluation platforms have certain eco-systems in place for SoC, power management, memory and other key components. The software provided by the silicon vendor would ensure basic functionality of the reference design. Based on these reference designs, OEMs would build a form factor prototype catering to their specific consumer electronics product vision. The OEM would then invest heavily on both Hardware and Software engineering resources to customize the hardware to fit the form factor of the Consumer Electronics Product Design and Development they envisioned. They would then negotiate with vendors and tweak certain components to bring out the final product to market.

Consumer Electronics Product Design

Consumer Electronics are all types of electrical devices, digital or old-school analog, that are intended for everyday use. They are usually small devices, can be mobile, stationary or wearable. Today, we have seen a radical change in the process of designing. Most fabless semi-conductor companies and ODMs operate in the heavily commoditized consumer electronics market space. Except for some big companies like Apple and Samsung, most OEMs are involved in only defining the specifications and functionality of products. The new age fabless semiconductor companies that dominate the consumer electronics or lifestyle electronics market today do not just stop at creating a reference design, but are also involved in the Consumer electronics product design which are nearly 80-90% completed.

Not only do they define the components that go into the product, they also bring together the entire eco-system of component makers required to for a Consumer Electronics product development. They negotiate and control the entire BOM of the product. Working with ODMs, they provide a product that can move into production at an accelerated time frame thereby removing the need for OEMs to have a large engineering team. So what does the OEM do? They define the product, differentiate it based on applications, UI, UX, industrial design, software features and spruce up the look and feel of the product as the core of the system is technically taken care of by the silicon vendor/ODM.

Role of ODMs in Consumer Electronics Product Development

• An ODM today is not just a design house, but is a “brain house” with the ability to envisage the entire product and bridge gaps between silicon vendors and OEMs. ODMs today need to have the ability to provide a ready-to-deploy Consumer Electronics product development and work with an EMS (Electronic Manufacturing Services) to provide a boxed product to the OEM.
• An ODM needs to continuously innovate and come up with consumer electronics product design and consumer electronics product development which can be converted to an end product with minimal time and effort thereby significantly shortening the design cycle, helping the OEM to introduce products to market at an accelerated rate.
• The base designs should have the flexibility to scale to different variants of the products while retaining the core functionality without compromising on the key features and still meet accelerated time-to-market schedules and quality.

This change in trend has brought about a revolution in both Consumer Electronics product development and design space where major brands especially in the emerging markets concentrate on brand building, distribution and marketing while leveraging the engineering strength of the silicon vendor and ODM. Their focus on in-house engineering is reduced drastically. This has resulted in significant cost savings which has been passed onto the customers in terms of a wide range of economical consumer electronics product development like phones, tablets and life style products which compete with brands like Apple and Samsung.

Click on this link to know more about Mistral’s Lifestyle and Consumer Electronics Product Design and Development services.

*Published in EE Times India

From smart phones to smart cars, technology is transforming virtually every aspect of our busy lives. Technologies not only make life easier while on the go, but new options are quickly simplifying life at home as well. Home automation is proving to be a top home improvement trend. Ninety million homes around the world will have some form of home automation solutions in place by 2017 according to a study.

2013 has been a fascinating year for home automation. We’ve seen growth at the luxury level, the do-it-yourself entry level and everything in between. While the costs and systems vary from home to home, most home owners are looking for ways to make their lives easier by making the system fit in with their lifestyles.

Home automation isn’t a one-size-fits-all product, so customization is the key. Call it the next frontier of personalized technology. The explosion of high-speed Internet and app-centric smart phones has made it easier for an owner to enjoy the benefits of home technology. This includes the luxury of controlling home systems from a Remote device (Smart phone or Tablet), round-the-clock security for the home even while the owner is miles away, maximum energy saving, thus cutting back on electricity bills and contributing to the creation of a greener earth. There are countless possibilities with home automation gateways and control systems. The connected home continues to evolve from basic connected home networks to a feature-rich collection of CE devices, broadband-enabled services, and multi-functional set-top boxes and home automation gateways.

There are several innovative Home Automation solutions being widely discussed in design labs across the globe. I’m sharing a quick overview of two of these technologies and their impact.

Digital Dog:  One of the ideas doing the rounds is that of a ‘Digital dog’. The ‘digital dog’ will be a home automation system focused on security. Security cameras, thermostats, door sensors and light modules that are wirelessly connected to the home’s high-speed Internet, allows you to control all of the home’s basic functions from any personal computing device, such as a tablet or smart phone. The ‘Digital dog’ will be in charge of monitoring the entire home. Imagine unlocking the front door from your office when the kids forget their keys or cranking up the living room’s heat during your commute home from work. For instance, if a postman is at your home for an important post delivery but you are at work, the Digital Dog system will interact with the postman, collecting his inputs which would be converted into a text message and sent to the user through the GSM network. This will help in keeping a tab on people who visit the house in the absence of the owner. Video monitoring of the house from anywhere can also be made possible through the Digital dog system.

Smart Building Automation Systems (SBAS): This is another booming trend in automation technology which can make meaningful reductions in the energy consumed by smaller commercial buildings or even residential projects. Lighting control systems use time and zone controls to turn off lights in any area of a building that is unoccupied, relying on occupancy sensors to indicate when people move from area to area and switching the lights on or off accordingly.

The air flow in Air Conditioning also can be controlled based on how many people are there in the floor through heat radiation sensors. Some of the other features include: Automated Entry and Exit automation to commercial office buildings using bio Metric sensor/smart cards and Correct Seat/Cubical occupation identification via Motion Detection Cameras and Access control Sensors. For example, if an employee enters a commercial office building, the system identifies and provides access to only his cabin or cubicle (PC, Light, AC, and Enter/Exit Door permissions).

With the speed of digital age, home automation technology is set to gleam in 2014. The Home automation world is shaping in two ways – by shrinking smart technologies into a smaller package and building a world that’s always connected.

*Published in EE Times India

Serial digital interface  is a standard for high quality lossless digital video transmission. In SDI Video Streaming designs, signals are uncompressed and self-synchronizing between the transmitter and the receiver.

SDI (Serial Digital Interface) is a professional video streaming signal that’s preferred in production environments because of its longer range (up to 300 feet) and reliability. SDI video streaming refers to the process of capturing, transmitting, and receiving high-quality video signals over a digital interface. SDI video streaming is a widely used method for transmitting uncompressed digital video signals over short distances. SDI technology is widely used in professional broadcasting environments due to its ability to deliver uncompressed video with low latency and high reliability and quality while ensuring compatibility with a wide range of professional equipment. SDI designs supports high-definition (HD) and ultra-high-definition (UHD) video formats, ensuring excellent image quality. SDI video streaming is preferred in production environments because of its longer range (up to 300 feet) and reliability. It is used for  transmission of uncompressed, unencrypted digital video signals or for packetized data.  SDI is used to connect together different pieces of equipment such as recorders, monitors, PCs and vision mixers. The greatest advantage of SDI technology is being able to transfer high- definition video signals without any loss of quality. This is due to the fact that the video is transferred in uncompressed format. A video network based on SDI Video can be easily put together with a readily available 75 ohm co-axial cable between a transmitter and a receiver. Because of these benefits, SDI Video streaming designs are rapidly becoming the leading video format for digital video transmission. Depending on the data rate there are different variants of SDI video streaming.

• SD-SDI carries NTSC/PAL video data with a data rate of 270Mbps
• 3G-SDI Video Streaming carries 1080p videos with a data rate of 2.97Gbps
• Dual – SDI carries two independent HD video streams in a single link. This results in a data rate of 2.97Gbps
• HD-SDI carries High Definition videos with a data rate of 1.485Gbps

Mistral offers Design Services for Video Streaming applications. One of our recent designs involved SDI Video Streaming (transmit and receive), both working at 3G data rate.

Solution Offered

The scope of this SDI Video Streaming project involved output of processed video streams to an SDI interface. We integrated a high-end video processor from TI along with a Spartan-6 FPGA from Xilinx to achieve the required SDI Video Streaming. SDI video streaming was implemented by integrating an SDI core inside the FPGA. The SDI Streaming core mainly uses two clocks, one a Reference clock and the other a pixel clock. The reference clock is a fixed LVDS clock input while the pixel clock is the one to which the parallel video data is synchronized from the processor. For proper functionality of an SDI Video Streaming application, the SDI core expects both data and reference clocks to be in complete phase synchronization.

SDI Video Streaming – Design Challenges

The Design Challenges that we faced in bringing up SDI-TX was w.r.t achieving synchronization between the processor’s pixel clock and the reference clock fed to the SDI core. To mitigate this, we added a FIFO in FPGA and used an internally generated clock inside the SDI core, as a FIFO read clock, with pixel clock being used as FIFO write clock. Although this resulted in phase synchronization, there were frequent underflow and overflow of FIFO due to minute frequency jitter among these clocks. To eliminate this, we had to find a common clock source for the pixel clock and the SDI core’s reference clock. To accomplish this, we connected a clock from FPGA, generated from SDI core reference clock, to an auxiliary clock input of the processor. Since the processor now derived the pixel clock using this auxiliary clock input, the FIFO overflow/ underflow issue got resolved.

Synchronization with the processor was not an issue for the SDI-RX path since the SDI-Core in itself generates pixel clock. The second major problem that we encountered was with respect to dual SDI video. To give DS-SDI output, SDI Video Streaming requires two parallel video inputs which are frame synchronized. In our design, the processor was giving out two parallel video outputs to FPGA and it was not possible to achieve frame synchronization at the source. Thus, the FPGA had to align the two video frames before routing it to SDI core. Here, we used a DDR3 connected to FPGA to achieve frame alignment. One of the incoming SDI Video inputs was written to DDR3 continuously beginning from the start of the frame, along with checking for start of frame in second video stream.

Once the start-of-frame in second video data is detected, this stream along with the first stream, is sent to SDI core for generation of DS-SDI. The data from the first stream is not live in the sense that a stored frame is being read back from DDR3. It has to be noted that the size of the DDR3 memory must be big enough to store one complete frame. Design Services for Video Streaming not only involves proper understanding of the SDI core architecture but also careful high-speed design. In addition, a major aspect to getting the interface to work flawlessly is a robust PCB design. This becomes extremely important since the data rate in SDI line can go up to 2.97Gbps. The main objective of PCB layout design is to achieve uniform impedance along the entire trace. This included careful selection of series components on the trace, trace width selection, trace separation for differential lines etc. Since our design involved both SDI video receive and transmit, we made sure that there was good enough separation between TX and RX circuits to avoid any interference.

*Published in EE Times India

Historically, power management for embedded devices was something that was mainly addressed at the hardware level. But in modern embedded systems, software has taken an increasing responsibility for power management. 

Power consumption by embedded devices is a key issue that needs to be addressed during any embedded product development. There is always a need to extend battery life and/or reduce the environmental impact of a system. Performance of embedded devices is highly dependent on adequate power management. Embedded devices are being packed with more and more features while the power budget has not changed much. Historically, power optimization and management in embedded devices was something that was mainly addressed at the hardware level. But, in modern embedded systems, software has taken an increasing responsibility for power optimization and management. This warrants effective power optimization and management using the available hardware and software resources; also referred to as hardware software co-design.

Implementation of Power Management Techniques

There are different levels at which Power Management can be implemented – from energy efficient peripherals and adaptive digital systems to power aware software programs. Device power optimization automatically reduces the amount of power used by individual devices or components when not in demand to perform some function. Disk drives, monitors, adapters, and even CPUs can provide this power-saving feature. The effect of device power optimization and management is transparent to the computer user. The overall system is still in operation, and is able to respond to requests for service from devices. Those devices are able to power up for full service within seconds when needed. Traditionally, power optimization and management has been about switching on/off devices as required for a particular use.

Recent developments in technologies have enabled chips with dynamic voltage and frequency scaling capabilities which is generally seen in microprocessors. The Power eco-system provides mechanisms starting from build time set ups to application software to make the implementation power sensitive. Starting from CPU scaling to set the core processor in a dynamically adaptive scaling based on load to a power infrastructure that can sense the power budget available and control application launches, these software setups provide considerable power management capabilities. While designing low power embedded devices, it is important to choose the right hardware components that can be put to power down modes and enable software to put hardware modules into different states of activity based on requirements for effective power handling. The hardware design inherently can bring in reliable real-time performance while software design bring in flexibility and configurability. Thereby, hardware software co-design increases the effectiveness of power management manifold.

An interesting example of power optimization and management in hardware software co-design is battery management in smart phones. When the battery level in your smart phone goes below certain levels, some of the services are automatically disabled. This is because of the hardware- software co-design in smart phones. Power management is done using the battery charger chip and a fuel gauge chip in hardware that performs  management and also provides battery statistics. Software then needs to consistently monitor battery statistics and helps policy manager to take decisions on power budget allocation to requests from applications. This is why certain requests like playing a video or audio clip will indicate a low battery signal.

When the device is inactive, the power policy manager automatically puts the device into power saving mode. Hardware software co-design is imperative for efficient power optimization and  management given the present day design challenges of time to market, cost, energy efficiency and ever increasing product features and design complexity. It comes with its own set of dynamism and complexities, which when handled with diligence and focus, yields great results. As a designer, hardware software co-design is a very powerful tool for efficient power management, to meet system demands in an optimized manner.

*Published in EE Times India

Virtualization is a term that has taken some time to internalize for me. My introduction with virtualization was as far back as 1992 when Windows 3.1 introduced the 386 enhanced mode, and the ability to run multiple DOS sessions with each session being independent of each other but having a separate 640KB space available to each virtual DOS session (can you believe we had operating systems that ran in 128KB?) .

My “play” machine is an Intel I7 based notebook with 8GB RAM, 1TB Hard disk, runs Windows 8 Pro and using the Hyper-V virtualizer runs for me a virtualized Windows 7, and Ubuntu 12.04 (64 bit) concurrently. My Ubuntu virtual environment quite often runs a virtual android device too (recursive virtualization?). Everything running at once, and I’ll be damned…it’s faster than my “work” notebook (Intel I5, 4GB) running my work operating system…OUTLOOK!

At the Consumer Electronics Show (CES 2013) earlier this month, the best demo that I saw was at the Ubuntu booth. They had a Samsung Galaxy S III phone, that when you plug it into a docking station (Display, keyboard, mouse) would run a complete Ubuntu desktop version concurrently with the Android that is already running on the phone.

There has been a lot of talk about Ubuntu’s other announcement of building a phone operating system but this one got my interest. Think of the possibilities, you can carry your entire computing environment around with you, and plug it into another device to give it context. So based on what you plug it into the device can morph to what you want !!

This was running a quad-core ARM Cortex-A9 architecture, and piggybacks off some modifications to the Android kernel. Newer devices launching this year are going to be running quad-core ARM Cortex A15 architecture, which has the ability to run each core separately, and you can actually run a separate operating system on each of the cores. I don’t really expect to see a virtualized IOS/Android/Windows phone anytime soon, but think about the performance that you can put in the background.

I got to discussing the demo and it’s possibilities with Pat McGowan, Director of Engineering and Product Strategy at Canonical (company the ships ubuntu). A lot of this is already available and can be used in your designs, the potential is huge…

The smartphone has become the window into the world around us. With features like these the smartphone will be able to acquire context, of where we are and what we want to do, and provide today what we only saw in the movies.

An ancient Chinese curse says “May you live in Interesting times…” We most definitely are

*Published in EE Times India

The ecosystem of chips for battery management and power optimization on feature–rich devices allows the designer to develop schemes to minimize power consumption while supporting power hungry features.

In general, power optimization seems opposed to product feature enhancement. Power optimization on feature–rich devices is built with high degree of focus on optimizing power consumption based on the use case. In the case of a mobile or a tablet the ecosystem is now quite well developed for power management and optimization; processor chips are made with lower micron fabrication, they operate at lower operating voltages and support many operating modes like (active, idle, sleep, deep sleep etc.), dynamic voltage and frequency scaling reduce processor power consumption. There are also power management schemes like AVS (Automatic Voltage Scaling) and DPS (Dynamic Power Scaling) to provide power scaling as required. The ecosystem of chips for battery management  and power management on feature–rich devices allows the designer to develop schemes to minimize power consumption while supporting power hungry features. This is now the standard approach to power optimization in feature rich devices, but what if you need to implement power optimization on a power hungry device?

Implementation of Power Optimization Techniques for Embedded Applications

There are Several techniques that can be deployed to to reduce active and static power consumption. Typically, the highest impact on power consumption is due to four key components – the processor, DDR memory, display and power design. Power management and optimization can be achieved by selecting the right system components and designing dynamic software architecture. In the instance I mentioned above, the ecosystem is well developed and optimized and it is a matter of porting the right schemes onto the existing devices. During the course of development, we came across an instance where the power consumption of an IP Camera for surveillance application had to be limited to a certain number. This involved power hungry ARM+DSP, FPGA and Video Encoder/Decoders. Given that the processor was not made for power optimized operation, we had to look at other areas to control and optimize power. The two main approaches adopted were 1) to choose the right chips with various modes (active, sleep, reset etc.) and 2) build intelligence into the electrical design to power on/off devices based on the use case of the product.

Given that the product had to operate in extended temperature grade; choice of the chips was not easy. Finally we chose video encoder/decoder chips that either have auto power down mode or external pin / register for putting the unit in power down mode. Certain chips like some of the video decoders did not have any of these, we had to control the reset and power to these chips. The FPGA chosen was from a low power family, we had an understanding that FPGA was a strong source of power consumption due to switching. This FPGA was built on a low-power 45nm, 9-metal copper layer, dual-oxide process technology for power management and optimization. Also during bit file generation, the tool offers options to optimize the design for low power consumption built on realizing the hardware with lower switching. The leakage currents on I/Os were reduced by putting the unused I/Os in tri-state modes.

The Power management and optimization schemes were developed within the FPGA based on the use case to put the various decoders into power down mode. The controls were driven from software on the processor. Electrical design was made such that the chip controls were accessible as GPIOs or I2C based registers, Software enabled/disabled the right chips on board. We observed that the DDR memory consumed a lot of power hence we had to find a mechanism to put the memory devices into dynamic auto management mode. The Power optimization on feature–rich devices schemes brought down the power consumption by certain levels but this was not sufficient. Given that we did not have an active power management infrastructure, the schemes designed had to be unilateral, static and could not be adaptive like in a mobile or tablet.

Then we started looking at the soft cores used on the FPGA and made them use-case sensitive. The cores were switched off / reset when not required. Then the bank voltages were selectively enabled / disabled for power optimization and further reduction of power consumption. Then we started looking at the leakage currents on pull up and pull down resistors in the design. This was used to very marginally reduce power consumption.  Presently we are looking at selectively using peripherals within the processor and map them to use cases.

Overall, we have achieved close to 90% power reduction or power optimization. The interesting part of this exercise was to come up with mechanisms of power management and optimization and saving in a power hungry design. We generally are given leeway on power consumption when the customer needs high-end product features. In this case most interestingly, the power constraint was because of system design. The power budget constraint of the IP camera we were designing was determined by the efficiency of the cooling mechanism, power budget of the system etc. We realized that power budget constraint derives its motivation from various system design aspects and not only the battery life. It was a refreshing alternate perspective into power optimization on feature–rich devices!

*Published in EE Times India

Till recently, connectivity used to be about being able to connect to the internet. Whether it was from a notebook, a tablet or a smartphone it was about being connected to the rest of the world. Over the past three years that has changed dramatically. Don’t get me wrong, being able to connect to the net is an important and fundamental requirement, but we also want to be able to connect to the things around us.

Where did I hear this recently…”Big things have small beginnings!”

I am a slave to my smartphone. It’s my alarm clock, my email, my entertainment, my reminders all rolled into one. I hate to admit this, but the first thing I do when I open my eyes in the morning is to reach out to my phone and check my email. I no longer have a newspaper subscription, I read the news on my phone. Twitter, facebook, linked in all there. But these are examples of being able to connect to the network and get stuff that is happening out there. I have a banking app, an app for what movies are going on, travel…all of it.

The phone goes everywhere with me. It’s a welcome weight in my pocket, and I cant put it down. One of the first thing that I check when I get up to go somewhere is do I have my phone with me. But then you already know all of this. Maybe you are a little like me when it comes to your phone too. This post is not about the ubiquity of the smartphone, but how things are changing…

From being my little window into the world, my smartphone has become my remote control to the world. I have a little device installed in my home, the Peel remote, it’s a little dongle that is plugged into my home network which talks over ZigBee to a small fruit that sits on my coffee table (actually I mounted it on one of my home theatre speakers). http://store.apple.com/us/product/H4714LL/A It’s a program guide application that tells me what’s playing on TV at any given time, and using the peel application I can say “Watch on TV”. That turns on my TV, my set top box, my home theatre. All from my smart phone.

I have an Apple TV; and all of my music is either on my home PC or on my phone. Connect to apple TV and voila I have my music. I’m watching a video on youtube and I think I want to show it to somebody, play it on my TV.

I’ve become a bit of a exercise fanatic, so I have two tools of the trade…a nike + ipod http://en.wikipedia.org/wiki/Nike%2BiPod. It tucks into my shoe (ok I cant afford the Nike + shoe and have a little pouch that attaches it to the top of my shoes!), and a heart rate monitor. It logs my steps, my speed (on the treadmill) and my heart rate. It tells me when I am pushing it too hard, and when I am pushing too little.

I hear that there is a device available that will track my sleep; a weighing machine that my phone can connect to to log my weight; a BP machine that will log my Blood Pressure to my phone, which will then log it online for me; I want them all! And I suspect I am not alone… I’m still waiting for a good watch that has the battery life and supports bluetooth connectivity so that when I get of the plane in a new place, new country, my watch is telling the correct time.

Do I really really need this? Maybe not. Am I happy to have this information with me? Definitely!

Are these small things…definitely, but there are big things too. Imagine a doctor walking up to a hospital bed and with a tablet in his hand (not the swallowing kind!) he/she has a view into everything connected to you. Your smart phone is your key to the door, your mobile wallet and a whole lot more.

I’m a techie, I like to use technology. I also have the pleasure of being part of building it in the first place. A few of the things I listed above, Mistral has been a part of!

Ah yes…I remember where I heard that quote I put in the beginning of this post, in the trailer for the movie Prometheus…Oops!

*Published in EE Times India