Smart Cities Archives - Page 485 of 568 - SmartCitiesElectronics.com

Infineon claims industry’s first true 1000A voltage regulator for AI

Tue, Mar 26, 2019

Adding to its high current system chipset portfolio, Infineon claims to offer the industry’s first 16-phase digital PWM multiphase controller.

The XDPE132G5C extends the existing portfolio which enables currents of 500 to 1000A and above for next generation CPUs, GPUs, FPGA and ASICs used in artificial intelligence (AI) and 5G datacomms applications.

It has been introduced as CPU current requirements increase to enable next-generation AI and networking workloads, DC/DC voltage regulators to deliver more than 500A to the load. The XDPE132G5C has a true 16-phase digital PWM engine and an improved advanced transient algorithm to address these high phase count requirements, says Infineon. The true active current sharing between phases enables a reliable, compact and cost-saving design, with no need for extra logic doubler ICs.

The XDPE132G5C offers fine V out setting in 0.625mV increments to meet the demands of ASICs and FPGAs of V out control in less than 1mV steps, seen in communication systems today. The XDPE132G5C also supports auto-restart for communications with options to reduce remote site maintenance following power or system glitches.

The XDPE132G5C is packaged in a 7.0 x 7.0mm 56-pin QFN to accommodate 16 phases. It employs a full digital and programmable load line and is PMBus 1.3/AVS-compliant.

Infineon advises that it can be paired with TDA21475, the thermally efficient integrated current sense power stage, to efficiently deliver over 1000A.

The 70A-rated TDA21475 power stage is housed in a 5.0 x 6.0mm package. It provides efficiency of more than 95 per cent. The exposed top significantly reduces the R th(j-top) from 19 degrees C/W in the over-moulded package to 1.6 degrees C/W. This removes heat from the top of the package, for voltage regulator power density and optimal phase count and footprint. The TDA21475 also offers smart over-current and over-voltage protection and delivers temperature and current information to the XDPE132G5C controller.

The company also offers the IR35223 true 10-phase PWM digital controller. This controller provides a cost-effective option for voltage regulation up to 500 A. The IR35223 is housed in a 6.0 x 6.0mm, 48-pin QFN package and provides advanced transient performance and telemetry features including PMBus 1.3/AVS bus compliance.

http://www.infineon.com/next-gen-processors

> Read More

Four-channel automotive PMIC is compact

Tue, Mar 26, 2019

Claimed to be the most efficient power supply available today, the MAX20049 is a power management IC (PMIC) for vehicle camera modules that has been released by Maxim Integrated.

It has been introduced to meet the demands of automotive camera modules which are becoming increasingly smaller. The MAX20049 integrates four power supplies into a small footprint. The MAX20049 offers options to support various output voltages while also providing fault mitigation by flagging faults and shifts in output voltages.

The four-channel MAX20049 measures 38mm² and is almost 30 per cent more compact than competitive solutions, claims Maxim Integrated and offer the highest efficiency among other quad-power power management ICs in its class. It has options to support modules that need various output voltages for different mixes of sensors and serialisers, enabling designers to make changes in layout or to fine-tune the IC to meet specific application requirements. Flexible sequencing and fixed output voltages support various image sensors. The MAX20049 provides fault mitigation to help flag faults and shifts in output voltages to ensure that the cameras are working as needed.

It has over-voltage protection, under-voltage lockout, external power good (PGOOD) signal and cycle-by-cycle current limit. And supports a wide voltage input range from 4.0 to 17V, enabling power-over-coax (POC), typically from 8.0 to 10V. There is also an option of using one of the buck converters as an intermediate supply for generating typical sensor and serialiser rails.

System efficiency at full load is 74 per cent, compared with 69 per cent efficiency for discrete ICs.

Spread spectrum and 2.2MHz switching frequency mitigates electromagnetic interference (EMI) to meet CISPR low-noise specifications.

The MAX20049 is available now and Maxim also offers the MAX20049EVKIT# evaluation kit and EE-Sim models.

This news story is brought to you by softei.com, the specialist site dedicated to delivering information about what’s new in the electronics industry, with daily news updates, new products and industry news.

To stay up-to-date, register to receive our weekly newsletters and keep informed of the latest technology news and new products from around the globe. Simply click this link to register here: softei.com

http://www.maximintegrated.com/products/MAX20049

> Read More

ML inference software library makes CNNs power-efficient

Tue, Mar 26, 2019

Machine learning (ML) inference software from Synopsys is optimised for low power IoT applications that use convolutional neural networks (CNNs) and recurrent neural networks (RNNs).

The embARC Machine Learning Inference (MLI) software library supports the energy-efficient Synopsys DesignWare ARC EM DSP and HS DSP processors. It is claimed to boost performance up to 16X for 2D convolution layers compared to unoptimised implementations and accelerates RNNs up to a factor of five for a range of topologies, including those built with long short-term memory (LSTM) cells.

The embARC Machine Learning Inference software library is designed to help developers create power-efficient neural network system on chip (SoC) designs incorporating Synopsys’ DesignWare ARC EM and HS DSP processor IP. The software library provides developers with optimised functions to implement neural network layer types, significantly reducing processor cycle counts for applications that require low power and area, such as voice detection, speech recognition, and sensor data processing, says Synopsys.

The embARC MLI software library is available through embARC.org, a dedicated website that provides software developers access to free and open source software, drivers, operating systems, and middleware supporting ARC processors.

The embARC MLI software library supports ARC EMxD and HS4xD processors and provides a set of kernels for effective inference of small- or mid-sized machine learning models. It enables the efficient implementation of operations such as convolutions, long short-term memory (LSTM) cells, pooling, activation functions such as rectified linear units (ReLU), and data routing operations, including padding, transposing, and concatenation, while reducing power and memory footprint.

Synopsys cites the example of low-power neural network benchmarks such as CIFAR-10 running on an ARC EM9D processor which can achieve up to a four fold reduction in cycle count compared to competitive processors in the same class, claims Synopsys. The MLI library provides an average of three to five times performance improvement across a wide range of neural network layers, such as depth-wise 2D convolution, fully connected, basic RNN cells, and LSTM cells with a maximum performance boost of up to 16X for 2D convolution layers.

Picture credit: iStock

https://www.synopsys.com

> Read More

Infineon claims industry’s first true 1000A voltage regulator for AI

Tue, Mar 26, 2019