Add to favorites

#Industry News

Autonomous Car Ingredients: Safety, Surveillance and Infotainment, Part Three

Autonomous Car Ingredients: Safety, Surveillance and Infotainment, Part Three

The Challenge

Image detection and processing needed for the smart camera applications has traditionally been difficult to achieve within the required size, power and cost constraints (see figure 1). An image is captured when light passes through the lens and falls on the image sensor pixel array. Each pixel converts the received amount of light into a corresponding number of electronics which are converted into a tiny voltage; the stronger the light, the more electrons that are generated, the higher the voltage. The voltage is then measured by an A/D converter and used in the image processing of the pixel array. This makes image processing of the sensor very susceptible to noise; both electrical and thermal.

Fig. 1: Challenges in Design

Electrical noise generated by camera electronics, in addition to the sensor itself, will lead to undesired variations of brightness (luminance noise) or color (chrominance noise) information. Such variations will directly impact the image quality. Chrominance noise usually appears as small, off-colored spots or specks in the image while luminance noise usually appears as small, dark spots that often look like film grain. The desire to shrink camera module size to fit today’s applications places constraints on not just component count and package dimensions but thermal power performance.

Higher temperatures, caused by reduced form factor enclosures, will increase significantly the amount of noise produced by an image sensor as well as reduce the reliability of the device. In addition to the camera module costs, cabling and connector harness selection is also a significant factor in overall system costs. Automotive applications must adhere to stringent EMI requirements, often resulting in the need for relatively expensive shielded or fibre cabling to reduce noise emissions.

Power Management

Power management technology is a critical element in IP camera applications. The image sensor requires a clean, low noise supply voltage to prevent banding, chroma and luminance noise from showing up in the picture. A high efficiency, low power PHY demands a high efficiency, low power source to match its performance. Meeting the demands of the input voltage supply is also crucial.

The power management system must operate efficiency at high conversion ratios when operating from the 48V input of a PoE supplied input. In automotive applications, it must withstand the high voltage surge requirements of the battery voltage. Micrel’s proprietary FetZilla and FleaFET MOSFET process technologies provide the optimum MOSFET performance for applications with 12V and 5V input requirements. The process allows MOSFETs with very low Rdson x Qg products, which enables our switching converters to operate at both high frequency and high efficiency.

Hyper-Light-Load (HLL) greatly reduces power supply input current at light loads, which allows high efficiency operation in low output power applications and also in applications that require a standby mode of operation. Space constraints have increased the challenge in managing heat inside the module so as not to degrade the CMOS sensor performance. While the main source of heat introduced is a function of the image processors, efficient power management can play a significant role by minimizing heat introduced and institute flexible, optimized power architecture.

Power management must be efficient and communicate with the processor as the power demand changes. A smart, efficient power solution ensures higher system reliability and increased performance while providing system optimization and easy adaptation to changes in CMOS sensor technology.

Copper pillar and “Hyper Light Load” are two of the Micrel technologies that meet these challenges. These two technologies fuse process and package technology into a single component in mixed signal integrated power devices to provide the best efficiency and performance. The improved process technology allows the necessary means to make it “Smart” by introducing increased digital content into the analog world.

Ripple Blocker active filter technology provides high frequency ripple attenuation for applications where switching noise cannot be tolerated. It provides up to 60 dB of attenuation at 5 MHz, cleaning up virtually all switching frequency ripple from upstream switching power supplies.

The High Efficiency, low drop out regulator (HELDO) family of converters provides the efficiency of a switching converter while benefiting from the high PSRR, low noise and fast transient response of an LDO. Copper pillar technology eliminates bonding wires and reduces parasitics, package size, thermal resistance and MOSFET Rdson (see figure 2). By reducing parasitics, it also lowers high frequency noise generated by switching converters. Lowering thermal resistance and reducing package size allows better power management integration with the sensor, compression and PHY components in the camera system.

Details

  • Riverside Center, 275 Grove St, Auburndale, MA 02466, USA
  • Sensors Online

    Keywords