EMBEDDED SYSTEMS Study material limayesir wordpress com Arjun

EMBEDDED SYSTEMS Study material: limayesir. wordpress. com Arjun Gour, Dr. S. S. Limaye ETC Department.

OBJECTIVE • To give sufficient background for understanding embedded systems design. • To give knowledge of RISC processor. • To understand connection of various peripherals with microcontroller based systems. • To study embedded system design aspects. SVPCET

OUTCOME • Students shall be able to design embedded based system. • Students shall be able to design embedded system based on RTOS and communication protocol. SVPCET

SYLLABUS UNIT -1 EMBEDDED SYSTEM INTRODUCTION History, Design Challenges, Optimizing design metrics, Time to Market, NRE And UNIT Cost, Application of embedded systems and recent trends in embedded system. UNIT-2 EMBEEDED SYSTEM ARCHITECTURE Hardware & Software Architecture, Processor Selection, Memory Architecture & IO devices, Interrupt Service Mechanism, Context Switching , Device Drivers. UNIT -3 ARM PROCESSOR ARM architecture & programming, RISC & CISC, ARM Organization, ARM programmer's model, Operating modes, Exception Handling, Nomenclature, Core extensions, Assembly language programming, Introduction to ARM Instruction Set. SVPCET

SYLLABUS UNIT -4 PROTOCOLS Bluetooth, IEEE 802. 11 and IEEE 802. 16, GPRS, MODBUS CAN, I 2 C and USB. UNIT-5 REAL TIME OS CONCEPTS Architecture of the kernel, Task scheduler, ISR, Semaphores (Shared data problem), Mailbox, Message queues, Pipes, Events, Timers, Memory Management. UNIT -6 CASE STUDY OF EMBEDDED SYSTEM BASED ON Communication (ROBOTIC OSCHESTRA), Automation (CHOCOLATE VENDING MACHINE, DIGITAL CAMERA), Security (SMART CARD), AUTOMOBILE (ADAPTIVE CRUISE CONTROL). SVPCET

History, Design Challenges, Optimizing design metrics, Time to Market, NRE and UNIT Cost, Application of embedded systems and recent trends in embedded system. Embedded System ? SVPCET

Definition • Embedded system is a system that has embedded software in a computer hardware which makes it a system dedicated for an application or specific part of an application or product or part of a larger system. • Embedded system is hidden. We don’t see it as computer. We only see the larger system. Embedded system may or may not have user interface. SVPCET

Difference between Embedded system & General purpose computing system SVPCET

Embedded System General Computing Systems ØBased on microcontroller ØBased on microprocessor ØMay or may not contain an operating system ØGeneral purpose operating system is present ØExecuting a specific set of applications. ØExecuting variety of applications ØApplication is preprogrammed. ØProgramme development on another computer ØApplication programmable by user ØProgramme development on same computer SVPCET

Embedded System General Computing Systems ØPerformance: • Less Power • Less Memory Fast execution ØPerformance: Faster is better. Don’t bother about memory or power ØTime requirement: • Highly critical ØTime requirement: Not critical • Built for harsh environment • May or may not have UI ØBuilt for normal environment ØMust have UI

Common characteristics of embedded systems • Real time and multirate – Reacts to events in deterministic time – Schedules functions in real time – Different operations take place at distinct rates, e. g. audio, video, data, network stream • Complex algorithms – Encryption, MP 3 encoding • Complex GUI or other UI • Dedicated functions (Runs same app for life) SVPCET

History of embedded systems • Before birth of computers, machine control was done by mechanical devices. • Even when electronic computer was invented, it was very expensive. Due to technological advances, the cost came down dramatically. So it became feasible to use it for machine control • Microcontrollers integrate CPU, memory and peripherals. Compact, cheap and reliable. • Today we find computers embedded into many household and industrial products.

Steam engine (1765) First industrial revolution. Computers were not known then. In old days, machine control was done by mechanical means. See how steam flow is directed to left and right side of piston in a steam engine.

4 stroke petrol engine (1876) Operation Suction Compression Power Exhaust V 1 off off on V 2 on off off Spark off on off Valves V 1 and V 2 are operated by CAM. Spark is operated by contact 6 o before top dead center.

OLD TV (1927) • Did not have remote control, timer child lock.

First electronic computer ENIAC 1946, Princeton, 18000 Vacuum tubes Started third industrial revolution

Birth of transistor • Shockley, Bardeen, Bratten Bell labs 1947

First embedded system • The first recognizable modern embedded system was Apollo Guidance Computer, developed by Charles Stark Draper at MIT Laboratory Instrumentation, early 1960 s. It was installed on Apollo spacecraft. It had both Command Module (CM) and Lunar Module (LM). Provided onboard computation to support spacecraft guidance, navigation and control. • Autonetics D-17 guidance computer was first mass produced embedded system.

Digital Equipment Corporation • DEC introduces the 12 bit PDP-8 minicomputer, which by 1973 was the best-selling computer in the world. PDP-8 s were probably the first computers "embedded" in instrumentation and other sorts of commercial systems due to its low cost (about $18 k in 1965 dollars, or $120 k today). • Ken Olsen (1977) said • There is no reason anyone would want a computer in their home • Went bankrupt in 1998.

8051, Father of modern embedded systems • Introduced in 1980 • First mass produced commercially successful product • 4 K ROM, 128 B RAM • UART, 2 timers, 4 GPIO • Modern versions – 64 K Flash, 2 K RAM – SPI, I 2 C, ADC

Modern embedded controllers • ATMEGA 328 P (Arduino uno) – 8 bits, 16 MHz, 2 KB RAM, 2 KB EEPROM 32 KB Flash • AT 91 SAM 3 X 8 E (Arduino due) ARM cortex M 3 – 32 bits, 80 MHz, 96 KB RAM, 512 KB Flash • STM 32 F 103 C 8 T 6 ARM – 32 bits, 72 MHz, 20 KB RAM, 64 KB Flash • TI SITARA AM 574 x – 4 cores ARM cortex A 15 1. 5 GHz, 2 C 66 DSP, 2 Cortx M 5 – GPU, GB ethernet, CAN, DDR 3 RAM IF • Broadcom BCM 2837 B 0, (Raspberry Pi) – 4 ARM 64 bit, Cortex A 53, 1. 4 GHz, GBE, DDR 2, Wi. Fi

Recent trends • • • Multi core processors (TI Sitara, Broadcomm) SOC (Wearable devices) Built in Wi. Fi (Node. MCU) Open source (Linux, Free. RTOS, Arduino) Low power Security (Embedded encryption) Authentication (Biometric) Device convergence (Mobile, Wearable) Internationalization (Multiple language & protocols) M 2 M comm (Water filter to service center)

Applications areas ØAutomotive electronics ØConsumer electronics ØAvionics (Aviation) ØFabrication equipment ØRailways ØSmart buildings ØTelecommunication ØLogistics ØHealth sector ØRobotics ØSecurity ØMilitary applications SVPCET

Examples of applications • • • Mobile phone Smart cards (Banking, security) Avionics (Missile, rockets, aeroplanes) Automotive (Cruise control, engine control) Consumer (Set top box, Digital camera) Computer peripherals (Network card, printer, graphic accelerator)

A longer list of embedded systems Anti-lock brakes Auto-focus cameras Automatic teller machines Automatic toll systems Automatic transmission Avionic systems Battery chargers Camcorders Cell phones Cell-phone base stations Cordless phones Cruise control Curbside check-in systems Digital cameras Disk drives Electronic card readers Electronic instruments Electronic toys/games Factory control Fax machines Fingerprint identifiers Home security systems Life-support systems Medical testing systems SVPCET Modems Speech recognizers MPEG decoders Stereo systems Network cards Teleconferencing Network systems switches/routers Televisions On-board Temperature navigation controllers Pagers Theft tracking Photocopiers systems Point-of-sale TV set-top boxes systems VCR’s, DVD Portable video players games Video game Printers consoles Satellite phones Video phones Scanners Washers and dryers Smart ovens dishwashers And the list goes on and on

Design metrics 1. 2. 3. Power dissipation: Important for battery operated devices – mobile, camera. Performance: Execution speed. Time elapsed between reception of signal at antenna to voice output on speaker should be less than. 1 s. Time taken by a Digital camera to capture process and store a 13 MP photo should be less than. 5 s. Process deadlines: There are several concurrent processes in an E. S. Mobile- audio processing, video processing, display refreshing, touch screen processing. Each process must finish computation within its deadline.

Design metrics contd. 4. User interface- GUI (Buttons, sliders, edit boxes) and VUI(Voice commands) 5. Size- Physical size(limited by end application) Memory size(RAM, flash in GB) No of million gates (ASIC or FPGA) 5. Engineering cost-One-time (Non recurring) cost of developing, debugging, testing, production jigs. 6. Manufacturing cost: Depreciation of machinery, PCB component placement, flow solder, harness, assembly, flashing, testing, packing, warranty, consumables and expenses.

Design metrics contd. 7. 8. 9. Flexibility-Ability to develop different versions of a product and advanced versions in future without incurring significant engineering cost. Prototype development time (Days or months) – Time required to develop a working prototype and in-house testing for functionality. Includes engineering time (design and simulation), prototype fabrication time and testing. UL approval, CE approval. Time to market (Days or months) – Time required to put the product in market after successful prototype. Brochures, advertizement, Raw material order, manufacturing jigs, pick and place programming.

Design metrics contd. 10. Safety. Equipment safety-Accidental fall to ground. Improper handling, over range inputs Theft – phone locking, tracing User safety- Automobile brake, engine, radiation hazards. 11. Ease of maintenance- Add hardware – Graphics accelerator to PC, memory stick to camera or phone Update hardware- Change 16 GB stick with 32 GB. Update software- Change OS, new version app Update data- new ring tone, wall paper, extend expiry date of smart card.

Design challenges 1. Amount and type of hardware needed • Minimize requirement of microprocessors, ASIPs, single purpose processors to achieve desired performance on design metrics: Power, cost (development and manufacturing) size etc. • Choose appropriate memory(SRAM, DRAM, internal and external Flash) peripheral devices (GPIO, timer, UART, SPI, LCD, LAN) 2. Minimizing power consumption • Reduce clock rate for non critical tasks or idle state • Disable cache for non critical tasks or idle state • Reduce voltage for non critical tasks or idle state • Introduce Wait (power down) mode and wakeup mechanism. e. g. mobile phone.

Design challenges contd. 3. Meeting process deadlines Must execute all tasks before their dead line considering memory, power, clock rate, cost. E. g. displaying a streaming video- must meet deadlines for MP 3 decoding. 4. Flexibility and upgradability Product must be flexible to allow different optional features. E. g. a digital camera basic model offers USB interface. Advanced models will offer Wi. Fi and Blue tooth. It should be possible to upgrade, i. e. introduce more advanced features in future. e. g. multi touch screen, voice commands, remote control

Design challenges contd. 5. Reliability Embedded systems are often used in critical applications like automotive, life support, industrial control. Reliability is important. Ensure reliability by proper design, Testing- Find and remove errors in software Verification- Ensure specific functions are correctly implemented Validation- System behaves as per the specifications Reliable hardware – Industrial grade components, burning in, IP 65 enclosure, EMI EMC, conformal coating, surge protection. Self test, watchdog timer

Optimizing design metrics • Cost of processor v/s process deadlines – Low cost processor has low clock rate and less word size. Requires more time to do same thing than expensive processor. Affects latency i. e. response time to events and ability to meet deadlines. • NRE cost v/s unit cost – SOC has very high NRE cost but low unit cost. • Size v/s NRE cost, performance, power – Reducing size requires VLSI in finer device geometry (e. g. 15 n. M instead of 22 n. M). NRE cost is high. Increasing size by parallel processing increases performance and reduces power. • Time to market v/s NRE cost – Using in house IDE tools and IP will reduce NRE cost but increase time to market. Using ready IP, it can be reduced.

Classification of embedded systems • Based on size • Small – Single 8/16 bit controller. Hardware and software complexity is less. No OS (Bare metal). IDE (Editor, assembeler/Compiler is used. Programming in C or assembly. Whole program fits in memory. • Medium – 16/32 bit controller, DSP, RISC. Use single purpose processors and IP (e. g. bus interfacing). Programming in C++/Java. RTOS used for multi threading. IDE contains source code control, simulator, debugger, JTAG emulator.

Classification of embedded systems contd • Based on size contd • Large (Sophisticated) – Enormous hardware and soft complexity – Use several IP, ASIP, FPGA, scalable processors – Constrained by processing speed – HW/SW codesign needed. E. g. encryption, DCT, TCPIP stack may be done by hardware.

Based on timing constraints • Hard real time system • Hard real time is a system that must perform tasks with the right deadlines. An example of a realtime hard drive system is a system that must open the valve within 30 milliseconds when the humidity of the air crosses a certain threshold. If the valve is not opened within 30 milliseconds it will cause havoc. The real-time hard system is often used as a controller for dedicated applications, having fixed time limits. Processing must be completed within defined constraints, or the system will fail.

Based on timing constraints contd • Soft real time system • Soft real time is a system that does not require deadlines. Example of soft realtime like DVD player, if given a command from the remote control it will experience delay for several milliseconds to run the command. This delay will not result in anything serious. Real-time soft systems have fewer hard time constraints, and do not support deadlines by using the deadline.

Revenues ($) Time-to-market: a demanding design metric Time (months) • Time required to develop a product to the point it can be sold to customers • Market window – Period during which the product would have highest sales • Average time-to-market constraint is about 8 months • Missing the window (start with a delay) will result in loss of revenue. • Need to balance time to market with NRE cost.

Losses due to delayed market entry Peak revenue Revenues ($) Peak revenue from delayed entry On-time Market fall Market rise Delayed D On-time entry Delayed entry W 2 W Time • Simplified revenue model • Product life = 2 W, peak at W • Time of market entry defines a triangle, representing market penetration • Triangle area equals revenue • Loss : The difference between the ontime and delayed triangle areas

Area = 1/2 * base * height For simplicity let H = W On-time RN = ½ * 2 W * H = W 2 Peak revenue Revenues ($) Peak revenue from delayed entry On-time Market fall Market rise Delayed D On-time entry Delayed entry W 2 W Time Delayed RD = ½ * (2 W – D) * (W-D) = ½ (2 W 2 -3 WD + D 2) Percentage revenue loss = ((On-time-Delayed)/On-time)*100 2 = (W 2 - W 2 + 3/2 WD – ½ D 2)/W 2 * 100 = D(3 W-D)/(2 W 2) *100 Ex. 2 W = 52 weeks, D = 4 weeks % loss = 4(3*26 -4) /(2 * 262) * 100 = 22% If D = 4 weeks, loss = 50%

Optimizing NRE and unit cost metrics Costs: – Unit cost: the monetary cost of manufacturing each copy of the system, excluding NRE cost – NRE cost (Non-Recurring Engineering cost): The one-time monetary cost of designing the system – total cost = NRE cost + unit cost * # of units – per-unit total cost= total cost / # of units = (NRE cost / # of units) + unit cost

Example – Consider NRE cost = $2000, unit cost = $100 – If we make 10 units, – total cost = $2000 + 10*$100 = $3000 – per-product cost = $2000/10 + $100 = $300 – If we make 100 units, – Total cost = $2000 + 100 * $100 = $12000 – Per-product cost = $2000/100 + $ 100 = $120

Compare technologies by costs -- best depends on quantity • Suppose we have 3 technologies to choose from – Technology A: NRE=$2, 000, unit=$100 – Technology B: NRE=$30, 000, unit=$30 – Technology C: NRE=$100, 000, unit=$2

A: TC = 2000 + 100 Q UC = 2000/Q + 100 B: TC = 30000 + 30 Q UC = 30000/Q + 30 C: TC = 2000 + 100 Q UC = 100000/Q + 2 Tech. A is best for volumes up to 600. Tech. B is best for volumes from 600 to 2600 Tech. C is best for volumes beyond 2600

The performance design metric Clock frequency and bus width • MIPS- Mega Instructions per second • Instr. cycle time is variable for CISC. Take weighted avg. • RISC takes one instruction in every clock cycle but sometimes the pipeline stalls, so effectively about 1. 5 cycles. • MIPS = Clock frequency / No. of clocks per instr. • MFLOPS- Mega floating ops per second • Important for number crunching applications (DSP, DIP) • Dhrystones (per second) • Benchmark program (Integer and string ops) • EEMBC (EDN Embedded Benchmark Consortium) • Set of benchmarks for various embedded application areas.

Measure of performance for camera Consumer electronics Ø Latency (response time) Time between task start and e. g. , Camera’s A and B process images in 0. 25 seconds Throughput Tasks per second, e. g. Camera A processes 4 images per second Throughput can be more than latency seems to imply due to concurrency, e. g. Camera B may process 8 images per second ) (by capturing a new image while previous image is being stored). This is possible with pipelining

Speedup is a common method of comparing the performance of two systems. Speedup of A over B = A’s performance / B’s performance