Applications in the Security & Surveillance Market

Agenda Infra-Red Camera CCTV Video Distribution Challenges & Market Trends AD5233, AD8668, AD7276, ADR125, ADG721 CCTV Video Distribution System Overview ADA4851-1, AD8131, AD8130, AD8113, ADV212, ADSP-BF548 The 2 signal chains in this brochure are for an infra-red camera and CCTV distribution. We’ll look at the trends & factors affecting the growth of security & surveillance markets, and the challenges that these 2 signal chains present.

Security & Surveillance Market Overview Market expected to increase 8.4% per year through 2008 Key Market Drivers Increased investment in developing regions Increased security concerns for tourism Lower system costs & increased system intelligence reduce manpower costs New IR sensors do not require cooling Enhanced image recognition algorithms The market for security systems and surveillance equipment is expected to increase by 8.4% per year through 2008. Aside from government programs that are designed to strengthen borders and inhibit terrorism, this growth is fueled by: Increased concern over how crime affects tourism Increasing investment in developing regions Ability of electronics to reduce manpower & increase effectiveness of systems.

IR Camera Block Diagram 1 3 2 IR Detector Typically has a pre-amp on output Cannot drive large current out DigiPot & VREF Corrects for DC bias on output due to: Bias voltage across detector Dark current (varies with temperature). PGA (amp + switched resistors) Allows 16 bits of dynamic range with 12-bit ADC. Looking at the block diagram for an Infra-Red camera, we see a detector and a signal conditioning circuit. The detector will be one pixel in a large array of pixels. A QVGA detector array will be 320 x 240 pixels. The detector array may have a pre-amp built in (shown as a buffer here), but cannot drive much of a load. In the past, IR detectors had to be cooled to get good performance. This made them expensive to design and power-hungry to use. Newer detectors do not need to be cooled to be reliable, but their output will change as the ambient temperature changes. Up to 80% of the infrared radiation that reaches the detector elements comes from the lens housing and other non-scene contributions. Any temperature change in the camera creates something similar to a DC level drift. To compensate for that thermal drift we will subtract a voltage we generate using a digital potentiometer and a voltage reference. The amplifier (item #3) is used to subtract the offset and switches between a gain of 1 and a gain of 4 before it drives the ADC input. We are using a PGA because the dynamic range of the IR detector output is very large. We need to stay sensitive to small changes in the IR information without being overdriven by very hot objects. By applying a programmable gain to the input we can get 16 bits of dynamic range with 12 bits of resolution. Standard definition video can only display 8-10 bits of resolution, but the additional resolution will be used by the DSP to enhance the image using edge detection algorithms. The last block in this signal chain is the ADC. We have a 12 bit ADC that operates at 3 MSPS. In order to get 30 frames per second we need to run the ADC at 2.3 MSPS (minimum). We calculate this by multiplying the pixel array size by the desired frame rate. The ADC that is chosen is a successive approximation ADC. For larger arrays the detectors may mux their output into 2 or more similar analog signal processing signal chains.

AD5233 — Quad 32-Position Nonvolatile Digital Potentiometer 10K, 50K & 100K Ohm End-to-End Terminal Resistances Extra programming commands simplify wiper setting Direct setting Increment/Decrement +/-6dB Increment/Decrement Readout of present setting 11 Bytes of Spare Nonvolatile User memory +2.7 to +5.5 Single-Supply or ±2.7V Dual Supply Operation 3-Wire SPI Compatible Serial Data Input & Output 700,000 erase/write cycle typical endurance with 100-year typical data retention The potentiometer shown in the signal chain is the AD5233. The AD5233 is a non-volatile digi-pot so it will maintain it’s last setting if it loses power momentarily. The controls for this include directly setting the value or sending increment/decrement commands. 32 positions are adequate for correcting offset due to ambient temperature drift. With an ambient operating temperature range from -40°F to +120°F (-40° to 52° C), that allows for a different calibration point every 5°F (approx 3° C).

AD5233 Quad NV Digital POT Applications: Mechanical Potentiometer Replacement in New Designs Sensor calibration LED Laser Intensity level & Extinction Ratio adjustment Gyro sensor adjustment Power supply voltage adjustment DIP Switch Setting Instrumentation - Gain & Offset calibration Digital potentiometers have many other applications. They can replace mechanical potentiometers in new designs – this saves assembly costs & eliminates the need for access points to turn the mechanical potentiometer. Some other uses include calibrating sensors (other than IR detectors) and adjusting the brightness or contrast in displays.

ADR12x Family - For Low Power Systems Precision, Low Voltage LDO Series References KEY BENEFITS Used with ADI’s 16-Bit Converters PB-Free Tiny Thin SOT-23 Pkg. Low Power Operation: Dropout Performance: 300mV Quiescent Current: < 85uA High Precision: A-Grade: 25ppm/oC B-Grade: 9ppm/oC Output Drive Capability: Source 5mA Sink 2mA 0.12% Initial Accuracy Package Temp Price TSOT-23-6 -40°C – 125°C $0.77 @ 1k In the block diagram shown, the digital potentiometer generates a programmable offset voltage based on a reference. The reference shown is the ADR125 This precision 5V reference can sink and source current and operates over a wide temperature range. Two of the reasons it was chosen for this block diagram are its small size and low temperature coefficient. We are using this reference to compensate for changes in the IR detector over a wide range of temperature, so having low temperature drift in our reference is important. All Options of the ADR39x Family ADR127 1.25V ADR121 2.5V ADR125 5.0V

2mV Precision Low Noise, CMOS Rail-Rail Output Op Amps iCMOS AD8668 Low Cost 2mV Precision Low Noise, CMOS Rail-Rail Output Op Amps KEY BENEFITS Low Cost, Low Noise 16V CMOS iCMOSTM Technology Best in Class for Performance vs. Price over the competition Excellent JFET alternative Small Package SOT-23 for single 5V Supply AD8665 2mV, 10 nV/rtHz $0.58 1ku TcVOS I BIAS Noise Bandwidth Vos Package Temp Price @ 1k The output of the IR detector and offset voltage (generated with the digipot) is buffered with a low noise, low offset op amp. The AD8668 is a single supply op amp that uses DigiTrim technology to improve precision while maintaining a low cost. DigiTrim is a way to trim the amplifier after it is packaged – this saves on cost. The AD8665/6/8 family also offer low bias current, low noise, and consume very little power. Another product in the DigiTrim family is the AD8661. AD8661 is the next generation of the DigiTrim family of CMOS amplifiers. The AD8661 extends ADI’s portfolio of precision amplifiers into the competitive, low-cost, CMOS market. The in-package trim technology allows for better precision at a competitive cost. The AD8661 family will extend the DigiTrim technology to a wider operating supply range: +/-8V, 5V to 16V. Extending the range will allow the advantages of DigiTrim to be used to address new customers and applications. Target Applications. Photo-diodes, Laser diode, EKG electrode sensing, power-line filtering 1 / 2 / 3 / 4-cell battery level sense, Point-of-load sense, Sense FET +/- 5V / 10V / 12V ADC & DAC buffering or filtering pressure, temperature, strain and other sensors Many of our customers are interested in using low cost precision parts in systems that operate on split supplies. They are interested in the precision, low input-bias, rail-to-rail input and output that is provided by the AD8605. By extending the operating range of these parts, we can fill the gap between our high voltage Bipolar parts and our CMOS DigiTrim parts. The new parts will have low input bias current and will help provide low cost options to our precision JFET amplifiers. 10 µV/°C max. 1 pA max. 10 nV/√Hz 4 MHz 2mV SOIC-8, SOT-23 -40°C to 125°C $ 0.58 AD8665 Single AD8666 Dual / AD8668 Quad SOT-23 SOT23 and SOIC Release: Q4 06 $0.58 @ 1kunits Dual 8-ld MSOP, SOIC, $0.78 @ 1kunits Quad 14-ld TSSOP, SOIC, $1.50 @ 1kunits 2.9 mm Ht: 1 mm

Switches for Consumer Products 1.65V – 5.5V Supply range Ron < 1 W Ron < 4 W ADG801/2 ADG849 ADG821/2/3 ADG884 3V, 5V ADG779 ADG701/2 ADG721/2/3 ADG736 ADG788 ADG819 ADG741/2 ADG787 ADG888 ADG859 Supply Range These are some other switches that are suitable for consumer products. Typically, a lower on resistance results in better flatness over frequency. The ADG721 has 2 single-pole, single-throw switches in a single package. In our block diagram we are using it to switch between a gain of 1 and a gain of 4 before we digitize the signal. While it is possible to use a switch with a lower on resistance, that typically results in a higher switch capacitance. That larger capacitance can increase settling time which would affect pixels that were digitized right after the gain was changed. The ADG721 offers the small size, low on resistance, and small capacitance that this type of application needs. ADG841/2 ADG839 ADG836 1.65V – 3.6V 1 x SPST 1 x SPDT 2 x SPST 2 x SPDT 4 x SPDT Switch Configuration

AD7276 – Highest Throughput in Smallest Package KEY BENEFITS High Throughput Up to 3 MSPS in TSOT Increased Throughput, Same Footprint Pin-for-pin compatible with the AD7476(AD7276) and AD7476A(AD7274) Ideal for Portable Applications Low Power: 12mW @ 3Msps Tiny Package: 9mm² No Pipeline delay: SAR architecture The ADC shown in this application is the AD7276, which is very similar to the AD7274 except that it derives the reference from the supply, it does not have a separate reference input pin. This ADC was chosen because it offers 12 bits of resolution and 3 MSPS with a SAR architecture. Our application is assuming a QVGA detector, which will have 320 by 240 pixels which will be 76,800 pixels in total. To sample each pixel 30 times per second (resulting in 30 frames per second) the ADC must run at 2.3 MSPS. The SAR ADC architecture is a good choice for this application because as the pixels are scanned there can be wide variations in temperature that can saturate the readings. For instance, if a car drives by the heat from the engine block will appear to white-out the image. So you will need to change the gain of the driver for subsequent pixels. If we used a pipelined architecture there would be several clock cycles until we had the pixel data we need to change the gain. Although standard video is only 8 or 10 bits of resolution, with IR images we typically want higher resolution because there will be some image processing used to improve the clarity of the image. By using a 12 bit converter and the PGA, we can deliver 12 bits of resolution and 16 bits of dynamic range for the image processor to work with. This makes it possible to see finer detail and enhance edges of objects for better image recognition. ** AD7276 – derives reference from supply AD7274 – separate reference input pin Resolution Throughput Channel No. Interface Power Supply Package 12-Bit 3 MSPS 1 Serial 2.35 V to 3.6 V 8-Lead TSOT AD7277, AD7278 AD7273, AD7274 As per AD7274 except 10/8-Bit resolution 3 MSPS 6-Lead TSOT package As per AD7276 except 10/12-bit Resolution, External Reference Input Pin

IR Camera Block Diagram 1 3 2 IR Detector Typically has a pre-amp on output Cannot drive large current out DigiPot & VREF Corrects for DC bias on output due to: Bias voltage across detector Dark current (varies with temperature). PGA (amp + switched resistors) Allows 16 bits of dynamic range with 12-bit ADC. This is one possible implementation of an IR camera, depending on the integration of the detector and the resolution requirements, a customer may have a very different block diagram. By understanding their signal processing needs you can find opportunities for other products from ADI.

CCTV (ADV212, ADA4851-1, AD8131, AD8130, AD8113, ADSP-BF548) The next block diagram we will review is for a CCTV (closed circuit television) system. CCTV is one way that video signals are routed from video and infra-red cameras to a central viewing area or video storage device.

AD8133 (Triple G = +2 Driver) & AD8143 (Triple Receiver) UTP Analog CCTV Systems Differential Drivers & Receivers and Cross Point Switches + VOCM VFB AD8131 DRIVER AD8130 RECEIVER 50W 100W VS- VS+ 750W 1.5kW 8 2 1 4 5 6 3 16 x 16 Cross Point Switch AD8113/4 Twisted Pair Cable CAT-5 DVRs Multiplexers Cameras Differential Drivers / Receivers (Ideal for UTP Systems) Drivers (cameras; camera modules): In a fixed gain of 2, no need for external resistors; unique internal feedback provides output gain and phase matching that are balanced to -60dB @ 10 MHz reducing radiated EMI & suppressing harmonic distortion. (Fully Differential; can be used single-ended-to-differential) Receivers (multiplexers; DVRs): high common mode reject ratio (CMRR) eliminates the effects of induced electromagnetic interference (EMI) over long cables; wide common mode input range maintains signal integrity in systems where ground potential is different between the drive and receive ends. (Differential to single-ended) Cross Point Switches (Multiplexers / DVRs that have analog multiplexing) 16 x 16 AD8113 (G = +2; 60MHz) AD8114 (G = +1; 225MHz) For RGB signals… AD8133 (Triple G = +2 Driver) & AD8143 (Triple Receiver)

ADV212 in CCTV Applications (FAQ) Why consider JPEG2000 when there is MPEG2/4 or H.264? More flexibility - encode at the highest quality (for detailed viewing) but only send portions that fit lower transmission bandwidth, resolution, frame rate, etc. To address HDD space concerns, the customer can reduce bit rate during scenes with no motion and increase quality when motion is detected Low latency - useful in remotely operating a camera Multiplexing multiple cameras - not possible with MPEG Regions of motion can be enhanced - not possible with MPEG Higher image quality - MPEG is not conducive toward this trend since it is 8-bits only and no still image capability Legal Issues – Some countries (Australia) do not allow MPEG still images as legal evidence (JPEG, on the other hand, has been used in the UK courts) What about JPEG2000 software decoders (for PC viewing)? Posted on our free FTP site: ftp://ftp.analog.com/pub/Digital_Imaging Still image quality – MPEG4 compression is an amalgamation of the I-frame plus the differences (Q-frames) Scalability – MPEG4 has some scalability but this must be determined during the encode process Latency is a key selling feature of the ADV202 – because the ADV202 uses a line-based transform, the compression is less than 1.5 fields worth; MPEG is up to 10 frames. The major reason why analog CCTV systems are still so prevalent is because of the latency involved in digital systems. Bit rate is always precise; the bit-rate jumps all over the place for MPEG, especially during zooming. JPEG compression bit-rate is image dependent. Ease of video editing – can’t easily edit with MPEG (temporal compression) In countries such as Australia, MPEG is unlikely to be admissible for use as legal evidence, compared to JPEG which has already been used successfully in a criminal prosecution in UK court proceedings.

When would MPEG2/4 be better than JPEG2000? MPEG2/4 is better with CCTV applications that have the following characteristics: Streaming only (one-way transmission) Bandwidth and storage space are the only concerns Mediocre image quality is okay Not a lot of movement in images In short – a “Digital VHS” type of application If the customer wants to do anything beyond “Digital VHS”, they should consider JPEG2000. For instance: Court accepted still images High quality still images Resolution, quality, and frame-rate scalability Conclusion: There are surveillance application that should use MPEG2/4; other surveillance applications would be better off with JPEG2000.

ADV212 Features Compliance to the JPEG2000 standard for still and motion pictures. SURF (Spatial Ultra-efficient Recursive Filtering) technology No external memory needed – competitive solutions need frame buffer memory Lower power – ~500mW of power dissipation Low Latency - <1.5 frames for encode or decode; MPEG solutions have up to 10 frames worth of latency Internal rate control is accurate to +/-5% MPEG or JPEG solutions are content dependent Tile/image sizes with widths up to 2048 pixels in 3-component 4:2:2 mode; 4096 pixels in single-component mode. Various standard video and microcontroller (16/32-bit) Interfaces for ease of use Flexible pixel interface supporting 8, 10, 12, 14, 16-bit Y, Cr, Cb pixels Compact 12x12 or 13 x 13 BGA package The ADV202 offers all of the advantages of JPEG2000 as well as our own design expertise.

JPEG2000 Competition TI offers various MPEG2/4 solutions for surveillance using TMS320DM642; however, each image is scaled to CIF and then compressed which degrades image quality Other MPEG vendors – Equator, LSI, Korean / Taiwanese vendors ADI offers their own DSP reference design; contact local ADI sales for more details Current DSPs (TI, etc.) are incapable of real-time video encoding / decoding of JPEG2000 JPEG2000 Competition Ricoh - D1 (720 x 480 @ 30frames/second) capable, but in a 420-ball, 35mm x 35mm package and also requires external memory. Sanyo – in-house JPEG2000 solution Sony - will use Sanyo OEM IP Vendors: Adelante, InSilicon, Amphion, others Ricoh - D1 (720 x 480 @ 30frames/second) capable, but it is also in a 420-ball, 35mm x 35mm package and also requires external memory as well. Sanyo – new solution to be introduced at the ISSCC (all of this needs to be verified – in addition, we are not aware whether there are any external memory requirements): 54MHz CLK 1440x1080 @30 frames/second Max is 4k x 2k (doesn't say how fast) Max CLK frequency is 104MHz Power is 400mW at 54MHz 0.18um, 9.2mm x 9.2mm, and has 18.9M transistors on it (this is probably false) Chin Fong Machine / Taiwan Univ. Chip: 720p at 30 frames/sec 3.2k x 3.2k max (doesn't say how fast) 0.25um, 5.2mm2 Current general DSPs are incapable of real-time video encoding / decoding of JPEG2000 Adelante – contains a combined hardwired ARM920 and DSP core solution with an integrated virtual prototype development agreement. We really haven’t seen this in the market, though it is the most “powerful” competitive device that we have seen as of today. InSilicon – Encoder and codec products available, but does not appear to be as powerful as the ADV202, with no scalability and no 9/7 wavelet. However, InSilicon’s large IP library may make them a contender in certain applications that require a lot of functionality. Cast (Alma-Tech) – a copy of the ADV-JP2000 but they have failed to add the 9/7 wavelet transform. The line-based transform would infringe on an ADI patent. Amphion – also claims line-based transforms, but their tile size limit of 128 x 128 makes the product almost worthless in some larger-image applications.

AD8113 Low Cost 16x16 Crosspoint Switch Array G=+2 Output Buffers KEY FEATURES Lowest-Cost 16x16 buffered switch array “Non-blocking” switch matrix Supports “broadcast” mode 80MHz -3dB Bandwidth Fully Buffered- Gain=+2 Drives 75W coaxial cable directly ±5 to ± 12V operation Handles a wide range of CM Voltages Switches composite, component video signals Several volume design-ins in CCTV applications!! Released to Production Shipping Now Several array sizes available- 8x8, 16x8 Differential I/O version in development!!

Low Cost Amplifiers for Coax Analog CCTV Systems ADA4851-1/ -2/ -4 Single, Dual, Quad ADA4860-1 Single ADA4861-3 Triple ADA4862-3 Triple; Fixed Gain = +2 Video Amplifier Low Cost Amps General purpose op amps that can be used for filtering and video buffering (Single: $0.55 US 1K units; dual: $0.79; triple $0.95; quad $1.09) Low cost amp can act as “sacrificial op amp” to save the more expensive crosspoint switch

ADSP-BDF548 High Performance Convergent Multimedia Blackfin Processor KEY FEATURES Blackfin® Processor Core with up to 600Mhz (1200 MMACS) performance Two independent DMA controllers Up to 260K bytes of on chip memory Connectivity: High Speed USB 2.0 OTG, Host DMA, UARTs, SPORTs, SPI, TWI, and CAN Multimedia: Multiple Enhanced Parallel Peripheral Interfaces (EPPI), and Pixel Compositor hardware accelerator Expansion: SD/SDIO and ATAPI-6 Controllers Synchronous memory interface for DDR / Mobile DDR connectivity Asynchronous memory interface for SRAM, EEPROM, NAND/NOR - Flash connectivity Blackfin® Processors embody a new breed of 16/32-bit embedded processor, ideally suited for products where a convergence of capabilities are necessary – multi-format audio, video, voice and image processing; multi-mode baseband and packet processing; control processing; and real-time security. The Blackfin's unique combination of software flexibility and scalability has gained it widespread adoption in convergent applications.

ADSP-BDF548 High Performance Convergent Multimedia Blackfin Processor Benefits OTP memory to enable users to implement private keys for secure access to program code. Human Interface: 18/24-bit LCD Controller, 32-bit Up/Down counter / Thumbwheel interface Lockbox™ Technology Hardware-enabled security for code protection The multiple Enhanced Parallel Peripheral Interfaces supports ITU-R BT.656 Video Formats and can drive 18/24-bit LCD displays A Hardware acceleration block, the Pixel Compositor, significantly reduces processor core overhead associated with software RGB-YUV color conversion and alpha blending. $18.66-21.35 @ 1K 533/600MHz Price Speed Package 400 MBGA Pb Free 17x17 mm Package 0.8 mm Ball Pitch VisualDSP++ for Blackfin Processors Easy-to-use integrated software development and debugging environment (IDDE). Efficiently manage projects from start to finish from within a single interface. To take a test drive, visit www.analog.com/aie/testdrive Or for more information on ADI Processors: www.analog.com/processors Key Applications IP Security and access control systems Biometric and Medical Systems Industrial & Instrumentation Automotive driver assistance/safety Telecommunications radio and switches Factory/building automation

Call to action Distribute booklets to your industrial customers Customers can order samples at special web page Maximum of 3 samples per part number, 12 samples per promotion. Additional Resources More Video Surveillance Solutions from Analog Devices Blackfin’s role in Infrared-Imaging Cameras Article on JPEG2000 systems in Design Line