0 3 5 ) . & 3€¦ · Engineering, Flex Power Modules Making Cables Safer By Adam Franczak, TME...

April 2020

Special Report: Industrial Applications (pg 21)

䤀渀昀漀爀洀愀琀椀漀渀琀漀倀漀眀攀爀夀漀甀爀䐀攀猀椀最渀猀一伀刀吀䠀䄀䴀䔀刀䤀䌀䄀

VIEWpoint

Need a Little Distraction?By Jason Lomberg, North American Editor, PSD

DESIGNtips

Step-Down Regulators for Automotive, Industrial Applications

By Dong Wang, Analog Devices, Inc

MARKETwatch

Automotive Applications without WheelsBy Kevin Parmenter, FAE Director, Taiwan Semiconductor NA

COVER STORY

GaN and 48 V – Where are We and

Where are We Going?

By Alex Lidow Ph.D., CEO and

Co-founder, Efficient Power Conversion

TECHNICAL FEATURES

Wide-bandgap Semis

SiC Semiconductors Propel Auto Industry Towards Energy EfficiencyBy John Palmour, CTO, Power & RF, Cree

Point-of-Load Converters

Powering Processors and FPGAs with the New TDK µPOLBy Philip Lechner, Avnet Abacus

SPECIAL REPORT:INDUSTRIAL APPLICATIONS

Switches: Critical to Modern

Industrial Sensors

By Craig Thomson, Director of Global Portfolio Management and Markets at C&K

FPGA Power System Management

By Pinkesh Sachdev, Product Applications Engineer, Analog Devices

Smart Connected Sensors Drive

Innovation on the Factory Floor

By Sergey Komarov, Business Development Engineer, TT Electronics

The Benefits of Hybrid Regulated

Ratio Topologies for Intermediate

Bus Converters

By Oscar Persson, Specialist R&D Engineering, Flex Power Modules

Making Cables Safer

By Adam Franczak, TME

Supercapacitors Deliver Reliable

Power for Warehouse Automation

By Jason Lee, Global product manager,

supercapacitors, Eaton

2

25

4

Highlighted Products News, Industry News and

more web-only content, to:

www.powersystemsdesign.com29

POWER SYSTEMS DESIGN 2020APRIL

1WWW.POWERSYSTEMSDESIGN.COM

32

38

42

13

8

COVER STORY

GaN and 48 V – Where are We and Where

are We Going? (pg 9)

22

9

16

13

FINALthought

The Industry’s Sustainability EffortsBy Jason Lomberg, North American Editor, PSD

Dilbert

48

48


© e

iCap

Happy withMLCC Downsizing?Check our menu.

#YOURCAPYOURSIZE

n Large portfolio from 0402 up to 2220

n Long term availability

n Detailed datasheets with all relevant

measurements and product data

n Sophisticated simulations available in

online platform REDEXPERT

MLCC Würth Elektronik offers a large portfolio of MLCC sizes up to 2220. While downsizing might be the

right choice for some applications, others require larger sizes of MLCCs for keeping the required

electrical performance, volumetric capacitance and DC bias behavior. Long term availability ex

stock. High quality samples free of charge make Würth Elektronik the perfect long-term partner

for your MLCC demands.

For further information, please visit: www.we-online.com/mlcc

PCIM Europe Hall 7 Booth 229

2

VIEWpoint

WWW.POWERSYSTEMSDESIGN.COM

POWER SYSTEMS DESIGN

Aren’t we in some kind of global pandemic? Let’s face it – we’ve all got bigger concerns than work, but you couldn’t tell from the overwhelming response to this month’s Industrial Applications topic.

Seriously, thanks for all your contributions! It’s not often I get the chance to thank our amazing industry partners, but this month’s flood of articles is humbling.

We open with a look at hybrid regulated ratio topologies for intermediate bus converters. These days, hybrid regulated ratio (HRR) converters combine the advantages of simple fixed-ratio converters with unregulated outputs and ones with full regulation and fixed outputs. Whereas, in the past, you had to choose one option over the other, HRR converters let you switch between each mode.

“To fully understand the value of HRR,” says Oscar Persson with Flex Power Modules, “it is necessary to compare the operation of HRR, fixed-ratio and fixed-output converters and consider each in relation to typical operating conditions.”

Meanwhile, Pinkesh Sachdev with Analog Devices discusses FPGA power system management. FPGAs have been a valuable commodity since the 1980s, when they originally evolved from programmable logic devices (PLDs), and they’re no less important today.

Indeed, as Pinkesh recounts, after Microsoft used FPGAs to speed up Bing in a 2013 pilot program, they expanded FPGA-equipped server usage to its cloud data centers.

Alex Lidow, with Efficient Power Conversion, delves into gallium nitride (GaN) and 48 V applications – specifically, where are we and where are we going?

We previously discussed on the PSDcast (podcast) how the demand for EVs and hybrids has fostered the adoption of the 48 V battery, and Alex takes it one step further – “Our reasoning to target 48 V applications was this major shift in topologies in automobiles and high-density computers that will translate into a large number of open design windows for new architectures.”

In the context of Industry 4.0 and the Internet of Things – where everything is connected to everything – TT Electronics' Sergey Komarov covers how smart connected sensors are driving innovation on the factory floor.

“Connected devices and smart equipment are streamlining manufacturing and supply chain processes on a global scale,” he says. “Smart sensors are at the heart of this movement.”

And finally, I’d like to shine the spotlight on a familiar (and endlessly fascinating) topic – wide band-gap semiconductors in the automobile. As John Palmour, with Cree, explains, “Automotive power electronics designers need to use advanced SiC technology to meet the high expectations on efficiency and power density,” and since SiC is an improvement over incumbent semiconductor technologies, and the aforementioned expectations will only grow, wide band-gap technologies will be inextricably linked to the auto.

Best Regards,Jason LombergNorth American Editor, PSD

Need a Little

Distraction?Power Systems Corporation 146 Charles Street Annapolis, MD 21401 USA Tel: +410.295.0177Fax: +510.217.3608 www.powersystemsdesign.com Editorial Director Jim Graham [email protected]

Editor - EuropeAlly [email protected]

Editor - North AmericaJason [email protected]

Editor - ChinaLiu [email protected]

Contributing Editors Kevin Parmenter, [email protected]

Publishing DirectorJulia [email protected]

Creative Director Chris [email protected]

Circulation Management Meghan [email protected]

Sales Team Marcus Plantenberg, [email protected]

Ruben Gomez, North America [email protected]

Registration of copyright: January 2004ISSN number: 1613-6365

Power Systems Corporation and Power Systems Design Magazine assume and hereby disclaim any liability to any person for any loss or damage by errors or ommissions in the material contained herein regardless of whether such errors result from negligence, accident or any other cause whatsoever.

Free Magazine Subscriptions, go to: www.powersystemsdesign.com

Volume 12, Issue 3


WWW.COILCRAFT.COM

®

The (lowest)Loss Leader

XGL Family power inductors feature the industry’s lowest DC resistance and extremely low AC losses for a wide range of DC-DC converters

Coilcraft XGL4020 Series molded power inductors are available in 14 inductance values from 0.11 to 8.2 µH, with current ratings up to 29 Amps. With up to 45% lower DCR than previous-generation prod-ucts, they are the most efficient power inductors available today! Their ultra-low DCR and higher Irms also allow XGL4020 inductors to operate much

cooler than other components. XGL4020 Series inductors are qualified to AEC-Q200 Grade 1 (with a maximum part temperature of 165°C) and have no thermal aging issues, making them ideal for automotive and other harsh environ-ment applications.

Download the datasheet and request free samples at www.coilcraft.com.

4

DESIGN t ips


Step-Down Regulators for AutoIndustrial ApplicationsBy: Dong Wang , Analog Devices, Inc.

The LT8609, LT8609A, LT8609B, and LT8609S are synchronous

monolithic step-down regulators that feature a wide 3 V to 42 V input range. This device family is optimized for applications requiring low EMI, high efficiency, and a small solution size—suitable for demanding automotive, industrial, computing, and communications applications. All regulators in this series have the same 2 A continuous, 3 A transient (

6

DESIGN t ips


Improved Thermal PerformanceEMI compliance is a concern in a number of environments, including automotive systems. With integrated MOSFETs, advanced process technology, and up to 2.2 MHz operation, all of the devices in this family can achieve a small solution size while satisfying the most stringent EMI standards. Spread spectrum frequency operation, which reduces EMI peaks, is available in all but the LT8609B. Furthermore, the LT8609S incorporates Silent Switcher 2 technology. Silent Switcher 2 devices feature integrated hot loop and warm loop caps to make EMI performance insensitive to board layout and the number of board layers. A board with fewer layers can be used to reduce manufacturing costs without sacrificing EMI and thermal performance.

Figure 2 shows that the LT8609S features the best peak and full load efficiency of the device family. Figure 3 and Figure 4 show a CISPR 25 EMI and thermal performance comparison of the Figure 1 solution on 2- and 4-layer boards.

ConclusionThe devices in the LT8609 family are easy to use monolithic step-down regulators with integrated power MOSFETs and built-in compensation circuits. They are optimized for applications with wide input voltage range and low EMI noise requirements. Low, 2.5 µA quiescent current and Burst Mode operation make them excellent battery-powered step-down converter solutions. A 200 kHz to 2.2 MHz switching frequency range make them suitable for most low power to micropower applications. Integrated MOSFETs and up to 2.2 MHz switching frequency capability minimize solution

feature spread spectrum frequency operation to reduce EMI peaks. The LT8609S displays the most impressive EMI performance in this family, based on its proprietary Silent Switcher® 2 technology, described below.

5.5 V to 42 V Input, Low EMI, High Efficiency 5 V, 2 A SupplyA 5.5 V to 42 V input to 5 V/2 A output power supply is shown in Figure 1. This solution features a 16-lead LT8609S regulator with a

2 MHz switching frequency. Only a few components are required for the complete solution, including inductor L1 and a few passive components. Figure 2 shows that this solution can achieve 92.9% peak efficiency.

Burst Mode Operation to Improve Light Load EfficiencyDuring light load operation and no-load standby mode, high efficiency and low idle current are important for battery-powered applications. The

Figure 2: Efficiency vs. load current for LT8609/LT8609A/LT8609S-based 12 VIN to 5 VOUT step-down converter

Figure 4: Thermal performance comparison between 2- and 4-layer boards for the circuit in Figure 1

Figure 3. CISPR 25 radiated EMI performance comparison between 2- and 4-layer boards for the circuit in Figure 1

LT8609, LT8609A, and LT8609S feature a low 2.5 µA quiescent current in Burst Mode® operation. During light load and no-load conditions, the switching frequency is gradually reduced, greatly reducing power loss while keeping the output voltage ripple relatively low. Figure 2 shows that light load efficiency remains above 85% while the power loss approaches zero at minimal loads.

High Switching Frequency with Ultralow EMI Emission and

Wolfspeed’s new 6.6 kWOn-Board Charger

delivers industry leading power density

The system delivers 54 W/in3 power density and exceeds 96.5% in peak efficiency with 67 kHz

switching frequency for CCM totem pole PFC and 150 kHz-300 kHz for a CLLC resonant converter.

visit wolfspeed.com/psd19 to learn more


size. CISPR 25 scanning results show excellent radiated EMI performance, compliant with the most stringent EMI standards. Silent Switcher 2 technology in the LT8609S makes its performance immune to layout and layer changes, which greatly reduces development and manufacturing costs.

Analog Deviceswww.analog.com

8

MARKETwatch


Our industry typically classifies power ap-plications that we can’t partition into

any other segment as “industrial electronics.” This includes ev-erything from energy-harvesting sensors to monitoring industrial machinery or a pipeline along its complete route, to the power levels needed to weld steel in robotic motion control systems and process control or other automation systems.

I often say that industrial appli-cations are automotive applica-tions without wheels, since using components that can pass an automotive qualification means that it can survive in industrial conditions. When designing industrial systems and selecting components it’s useful to ask your supplier if the parts have passed an AEC-Q qualification test. You don’t need the PPAP process and paperwork yet hav-ing the qualification test showing that the component has passed this testing is key. Using this best practice testing on every device – automotive or not – results in no-compromise assurances.

Industrial applications must be

Automotive Applications without WheelsBy: Kevin Parmenter, FAE Director, Taiwan Semiconductor NA

able to survive abuse by end us-ers, not to mention the environ-mental extremes of temperature, humidity, electrical overstress, shock and vibration, to name just a few. Thus, what is called for are designs backed by component selection and testing that goes beyond what is required. De-signing industrial applications to meet not just environmental requirements, but those intended for automotive and medical – even if your application doesn’t need it.

Many industrial applications are used in outdoor and -or harsh en-vironmental conditions, such as an aluminum smelting plant and can be exposed to -55C to +71C ambient temperatures. Systems are subject to constant electrical onslaughts, including surges and transients affecting I/O systems and power input and output lines, as well as back-EMF from motors or other inductive loads. Moreover, industrial applications must withstand unsophisticated users. A common mistake, for ex-ample, is accidently wiring a 240 VAC line to the thermocouple input pin, fixing it, and then ex-pecting it to work properly. Or ac-cidently wiring a 48-volt DC line

backwards and, again, thinking it won’t have issues.

When selecting a sub system, like a power supply, for your industri-al application, ask if it can meet the MIL STD 810G environmental testing standard. Additionally, ask if it can meet the UL508A standard for industrial control panels. Perhaps in your applica-tion you don’t have to meet this standard, but there are power supplies available that do. In the same way, make your power supply meets the new UL62368-1 for safety, even ask if it meets the medical safety standard UL60601 for good measure. After all, nobody ever got in trouble for too much safety. And for EMC, if it can meet the 60601-4th edi-tion standard, you’ll have even more low-cost insurance that the power supply likely will not interfere with, or be interfered by, something else in the system. Many commercial supplies offer these included and take advan-tage of them all if you can. My motto is too many certifications is just about right.

PSDwww.powersystemsdesign.com

GaN and 48 V – Where are We and Where are We Going?

By: Alex Lidow Ph.D., CEO and Co-founder, Efficient Power Conversion

WLP products are auto certification, been in orbit, been medically inserted into humans, and have saved a lot of energy

Three years ago, the cost of making medium voltage eGaN FETs fell below the cost

of equivalently rated power MOSFETs. At that time EPC decided to use the performance and cost advantages of eGaN FETs to aggressively pursue applications with input, or output, voltage around 48 V. Specifically, automotive and computer applications is where 48 V conversion is becoming the new architecture, the new standard for power systems.

Our reasoning to target 48 V applications was this major shift in topologies in automobiles and high-density computers that will translate into a large number of open design windows for new architectures. Having open design windows, designers will consider significant changes in their designs and aggressively evaluate new higher performance component technologies.

Positive results have followed − the increasing adoption of eGaN FETs and ICs in 48 V applications has

been very gratifying. High volumes have already been realized in high density computing, and many new automotive designs designs are in process that will cause GaN product volumes to approach that of their aging ancestor, the silicon MOSFET. A third application , motor drive, is also beginning to take advantage of GaN high performance and low cost with new, innovative topologies.

To demonstrate how 48 V conversion is accelerating GaN’s adoption, we will first consider computers, then automotive applications.

48 V Computer Applications with GaNMany server designs are switching from having rack-based 48 V isolated and regulated DC-DC converters that convert to 12 V (see Figure 1), to non-isolated, unregulated 48 V DC-DC converters mounted on the server boards.

Space on a server board is extremely precious, and servers – especially those designed for performing high-end computation functions such as Artificial Intelligence (AI) – are demanding more and more

Figure 1: Si MOSFET-based architecture for providing AC power to the CPU or GPU

COVER STORY


1110

COVER STORY

WWW.POWERSYSTEMSDESIGN.COM WWW.POWERSYSTEMSDESIGN.COM


even smaller and with improving performance. In March 2020, EPC launched the ePower Stage IC that combines power FETs in a half-bridge topology with their own drivers, level shift, and input logic onto a single GaN-on-Si chip (see Figure 4). The surprising result was that this tiny chip outperforms discrete

eGaN devices with the same nominal specifications when driven with a conventional silicon-based driver IC (see Figure 5).

There are three reasons for this superior performance; (1) reduced power loop and gate loop parasitic inductance, (2) driver devices are perfectly matched to their power

FETs and, (3) improved thermal management due to the ballasting effect that comes from putting two devices onto the same substrate.

In the near future, we expect that ePower Stage devices will be used in the primary side of LLC converters such as shown in Figure 2, or in buck converters such as shown in Figure 6. This tiny power stage is 35% smaller than the best layout possible with discrete eGaN FETs and a silicon-based driver IC. It has all the components of a buck conveter with the exception of the controller.

48 V Automotive Applications Using GaNIn the automotive world, there are similar shifts in power architecture underway. As the world shifts from pure internal combustion engines (ICE) to mild hybrids and plug-in-hybrids, many of the functions that were mechanically driven by the engine are converting to electric

Figure 4: EPC2152 ePowerTM Stage integrated circuit combines a monolithic half bridge with drivers, a level shifter, a synchronous bootstrp circuit, and input logic onto a single GaN-on-Si chip

Figure 5: The green lines show the efficiency of the ePower Stage IC at 1 MHz (solid line) and 2.5 MHz (dotted line). The blue lines represent the efficiency of the same circuit made with discrete eGaN FETs and a silicon-based driver IC. Note the black X that shows the best performance by a MOSFET-based converter at 1 MHz

Figure 6: This EPC9207 48 V – 12 - 5 V, 12.5 A, power stage measures 14 mm x 9 mm x 4 mm and uses the EPC2152 ePower Stage IC. It has all the components, except the controller, needed to create a buck converter

power. Simultaneously, GPU’s are becoming quite common in AI systems. These GPU’s are putting additional demands on power conversion requirements, where voltages as low as 0.6 V are required under certain operating conditions.

These stringent power requirements have necessitated changes in basic power conversion architecture. First, with board space at a premium, high power

density architectures becomes the holy grail for power systems designers. eGaN® FETs from EPC have been in production for 10 years, and the 5th generation devices are half the size of

their 4th generation predecessors, are twice as fast, and are priced comparably with MOSFETs.

For illustration of GaN’s superior performance, shown in Figure 2 is a comparison between a traditional 300 W isolated 48 V – 12 V silicon-based converter and an ultra-small 300 W, 48 V – 5 V LLC using eGaN FETs. In addition to a significant size reduction, the GaN-based LLC has 30% lower power losses.

Lowering the intermediate bus voltage from 12 V to 6 V (or 5 V) is a second shift that is rapidly occurring in server architecture. This shift allows a change in technology from power MOSFETs on the secondary side of the DC-DC converter to higher density BCDMOS ICs that include the power MOSFET and driver electronics.

In addition, the point-of-load (POL) converters that used to be 12 VIN to 1 VOUT and relied upon trench MOSFETs are now shifting to 6 V to 0.6 V POL converters using monolithic BCDMOS with a single silicon IC. These BCDMOS POLs can go to higher frequencies, making the POLs smaller and able to be placed much closer to the GPU or CPU. This reduction in distance can reduce the resistance between the POL and the GPU/CPU by as much as 350 µΩ. At 1000 A, that is a 350 W reduction in losses!

Additionally, the AC-DC power supply that remains on the server rack is also converting rapidly to GaN FETs or GaN ICs on both the primary side and the 48 V secondary side. This new architecture, and the technologies used for power conversion, is shown in figure 3.

Note: There are no silicon power MOSFETs in this new architecture.

Introduction of GaN ICs Reduces Size and Improves Performance GaN devices are rapidly becoming

Figure 3: Emerging architecture for providing AC power to the CPU or GPU. Note, MOSFETs have been replaced by GaN devices or BCDMOS ICs, and the 12 V bus voltage is now 6 V (or 5 V)

Figure 2: (Top) 300 W 48 V – 12 V 1/8 Brick DC-DC converter using Si MOSFETs. (Bottom) 300 W, 48 V – 5 V LLC using eGaN FETs on the primary side

12

COVER STORY


power. As such, electrical power requirements are growing rapidly from sub-1 kW to 2 – 8 kW. To handle the high current needed at these power levels would require very thick and expensive wires in the wiring harnesses. The automotive manufacturers are therefore planning to shift to a 48 V bus (see Figure 7). 48 V is a logical operating point because the diameter of the copper wire is reduced by a factor of 16, and 48 V is safely below the 60 V limit above which extra isolation layers would need to be added to electrical systems. As was the case for high density computing, GaN is the best

Figure 7: IHS Automotive predicts 11 million vehicles will have 48 V primary distribution busses by 2025

Figure 8: Shown is a comparison between a 3 kW MOSFET-based 48 V – 14 V bi-directional buck/boost converter (top) and the same power output design using EPC2206 automotive qualified eGaN FETs (bottom)

Figure 9: Cost comparison between MOSFET design and eGaN FET design of a 3 kW, 48 V – 14 V bi-directional buck converter for mild hybrids

choice for the power devices. An examble of this is shown in Figure 8. This is a comparison of a 3 kW, 48 – 14 V bi-directional converter using power MOSFETs and the same output power using eGaN FETs. By using eGaN FETs, the frequency can be doubled while still improving conversion efficiency and reducing the design from five phases to four phases.

The reduced cost of the eGaN FET- based design is shown in Figure 9. Higher efficiency, smaller size, less weight, and lower cost are compelling reasons for GaN devices to dominate this automotive application.

ConclusionGaN devices have over 10 years of actual customer experience in the field, and the volumes used in high-density computing and automotive applications have grown exponentially, bringing down the cost and increasing power system design engineers’ confidence. Computers and automobiles are just the beginning. Motor drives for robots, drones, and industrial automation are the next major markets poised to discard aging power MOSFETs and adopt smaller, faster, and lower cost GaN FETs and ICs.

Efficient Power Conversion (EPC)www.epc-co.com

SiC Semiconductors Propel Auto Industry Towards Energy Efficiency

By: John Palmour, CTO, Power & RF, Cree

In the race to build the premier electric vehicle, wide bandgap semiconductors are the sustainable key

It’s no secret that the auto-motive industry is striving for "zero emissions" trans-portation, which means

manufacturers are rapidly ramp-ing up their electrification pro-grams. Most OEMs target 2025 for significant sales volumes of battery and hybrid electric ve-hicles (BEV and HEV). To meet customer expectations on range, charge-time and performance, and to approach the critical adoption of emission-free cars, these EVs require power electron-ic devices capable of efficient and effective operation at elevated

temperatures. As a result, power modules are being developed us-ing wide bandgap (WBG) silicon carbide technology.

With global pressure to decrease energy consumption, increase efficiency and reduce greenhouse gases, CO2 emission reduction has become key in every layer of the value chain. Power electron-ics play an essential role in these demanding challenges. The push for higher efficiency and smaller more compact power manage-ment is driving innovation utiliz-ing breakthrough silicon carbide

technology. These advanced sili-con carbide module technologies will enable higher performance in traction inverters for trains, HVDC for power transmission and distribution, solar and wind inverters, energy storage, and solid-state transformers. For the EV market in particular, this means longer driving distances and faster charging times using the same size battery.

Silicon carbide set to dominate innovative automotive power ap-plicationsAutomotive OEMs are bringing

Figure 1: The above graph shows the strong growth opportunity at hand as emission standards drive a significant shift to battery electric vehicles. (Source: Yole and Cree estimates)


WIDE-BANDGAP SEMIS

14

WIDE-BANDGAP SEMIS


Working toward an improved WBG future Given the demands of the EV market, companies are making significant strides in the WBG semiconductor space. Notably, Wolfspeed, a Cree Company and a leader in silicon carbide power products, has developed a 1200V 13 mohm of continuous drain current at a case temperature of 25°C. Silicon carbide enables the reduction of EV drive-train inverter losses by 78. This efficiency improvement offers designers new options in terms of range, battery usage, packaging and vehicle design, while ensuring reliability through lower thermal stress.

Wolfspeed´s commitment to silicon carbide devices allows the company to offer industry-leading silicon carbide MOSFETs and silicon carbide diodes

for industrial and automotive applications. The company is currently investing heavily in the expansion of its silicon carbide production by up to 30-times (between 2017 Q1 to 2024) with the creation of the world’s largest silicon carbide fabrication facility in Marcy, NY and a mega materials factory at its U.S. campus headquarters in Durham, N.C. The plan delivers additional capacity for its industry-leading Wolfspeed silicon carbide business with a brand new, state-of-the-art, automotive-qualified 200mm power and RF wafer fabrication facility.

WBG semiconductors in 2020 & beyond The growth is primarily based on the fact that the future of silicon carbide in EVs is very promising. It’s clear that with the need for more range and power, as well as faster, more efficient charging, silicon carbide has the physical properties that best meet these needs. As we start a new decade, we are already seeing the demand for its use in EV development skyrocket along with an array of other industries. Those trends don’t show any signs of slowing down and the years ahead promise new waves of innovation as those industries improve their understanding of the technology and its benefits.

Creewww.cree.com

Figure 3: A Wolfspeed Silicon Carbide MOSFET

an increasing number of EV mod-els to the market, but due to their comparably limited range, accep-tance by the end user remains subdued. Aerodynamic design, and lighter materials, do help, but are not enough. Automotive power electronics designers need to use advanced SiC technology to meet the high expectations on efficiency and power density.New silicon carbide components are an improvement over incum-bent semiconductor technolo-gies such as silicon (Si) metal-oxide semiconductor field-effect transistors (MOSFETs) and insulated-gate bipolar transistors (IGBTs) by offering much lower losses, higher switching frequen-cies, higher operating tempera-ture, and robustness in harsh environments. They are particu-larly useful as the industry moves toward higher capacity batteries that operating at high bus voltag-es up to 800V with shorter charg-ing times and overall reduced losses.

Silicon carbide semiconductors hold great prom-ise to significantly outperform and replace traditional Si components. These new compo-nents have currently reached a level of maturity that has allowed it to rapidly gain ground over Si for demanding applications, such as traction power

electronics in EVs and HEVs.

Automotive market driving silicon carbide growthWhere silicon has a breakdown electric field of 0.3MV/cm, sili-con carbide can withstand up to 2.8MV/cm, and its internal resis-tance can be 100 times less than an equivalent silicon MOSFET. As a result, applications can handle the same level of current using a smaller, faster device, result-ing in smaller and more efficient systems.Consequently, silicon carbide in-verters allow for smaller cooling solutions and greater thermal-management design flexibility due to significantly less heat generation. Also, silicon car-bide inverters deliver a four-fold increase in drive frequency over conventional technology with no efficiency penalty in step-up converter applications, unlock-ing significant improvements in power output and density by allowing for smaller and lighter

peripheral components.

The increased efficiency of silicon carbide-based power semicon-ductors also increases the range for electric vehicles in contrast to standard silicon technology. Due to high battery costs, the efficient electric drive represents an enor-mous growth potential for the foreseeable future. In particular, silicon carbide technology in con-junction with the 800V vehicle electrical system voltage makes a significant contribution to further increasing efficiency which can be used to increase range or de-crease the amount of expensive batteries needed to deliver the same range.

Advantageous properties of WBG semiconductorsDoubling or tripling the bandgap in comparison to silicon means that silicon carbide devices can tolerate much higher voltages and electric. As a result, the on-resistance of a SiC MOSFET is far lower. This allows the use of a MOSFET, a very high-speed de-vice, in these high voltage appli-cations rather than a slow, lossy silicon IGBT device.

Overall, for a wide range of power applications (not just automotive), the capabilities of silicon carbide make it possible to reduce weight, volume, and life-cycle costs. In turn, dramatic energy savings can occur, and these applications are far more powerful than they were before.

Figure 2: A Wolfspeed Silicon Carbide Power Module

www.powersystemsdesign.com

THE BRIGHT SOURCEFOR ALL THINGS POWER.

16

POINT-OF-LOAD CONVERTERS


Powering Processors and FPGAs with the New TDK µPOL

By: Philip Lechner, Avnet Abacus

A new ‘plug and play’ solution to power complex ICs can space and design time without compromising performance.

Minimising power consumption is a high priority and in digital devices, it

is directly related to voltage swing and data rates. Consequently, supply voltages have steadily reduced from 5V to sub-1V with sub-sections of devices having separate supply rails: processor cores between 0.56V to 0.88V on next generation 7nm technology, a similar voltage for block RAM, 1.35V or 1.5V for DDR memory, 1.5V for auxiliary functions and 3.3V and 5V for legacy I/O. Further SoC integration is adding additional voltage requirements for DSP, codecs and video/graphics processing and often multiple, similar voltages need to be generated separately, to avoid interactions, and so they can be individually enabled and power sequenced.

Low voltages mean high currents for a given power, with some SoCs, FPGAs and processor cores now drawing more than 200A. Other rails might be from 1A to 25A for memory, SERDES, and I/O voltages. These rails are derived from Point of Load (PoL)

regulators from a higher bus voltage, typically 5V or 12V, where the current is much lower for the power drawn. System designers will generate a power ‘tree’ diagram to summarise the arrangement (Figure 1).

Putting PoL converters close to devices like FPGAs poses problems though; a typical FPGA such as in the Xilinx UltraScale range has more than 1000 BGA connections and is surrounded by supporting devices for I/O as well as a substantial heat sink. Space is at a premium and the PoLs have to fit in where they can, especially in small form factor applications such as FPGA SoM cards and PCIe cards. The ideal PoL is therefore highly efficient and small without needing heatsinking, and low profile, so that it can be back-side mounted or even fitted under other components such as board-mounted piggybacks or

sub-boards. Of course, it must also provide a precise voltage under static and dynamic load conditions, have output control for shutdown and sequencing and be fully protected against shorts, over- and under-voltages, and over-temperature stress. Modern PoL devices will also have communication features, typically over a I2C bus using PMBus commands, to set up device characteristics, as a one-time adjustment for fault detection thresholds for example, or even ‘on-the-fly’ for dynamic control of output voltage for energy saving. The bus can also signal PoL conditions such as load current, temperature and status.

Figure 1: A typical ‘power tree’ for a Xilinx Zynq 7 device

PoL basicsFigure 2 shows a basic PoL con-verter. Q1 and Q2, switch alter-nately to supply load current and ‘charge’ inductor L1 in Q1 on-period and discharge its energy through Q2 to the load during Q1 off-period. Controller IC1 samples the output and drives Q1 and Q2 with pulse-width modulated sig-nals to achieve regulation of the output voltage in conjunction with L1 and C2. IC1 also monitors and

detects abnormal conditions and integrates the communications function.

Component se-lection is a trade-off; L1, C1 and C2 can be smaller

at higher switching frequencies, but this produces more dynamic losses in Q1 and Q2, requiring a bigger part to dissipate the heat. Advances in MOSFET technology, however, allow operation in the MHz range today.

PoL designs - the ‘old’ wayComponents in the PoL shown in Figure 2 are discrete and could be placed on the motherboard. However, PoL design is a particular

skill and it is very difficult to achieve optimum performance when the PoL is mixed in with other motherboard components, also consuming precious motherboard space.

Modular PoLs have been available, offering the advantage of a single component to source, saving the customer design and qualification overheads, as well as procurement, stocking and placement of multiple discrete components. However, older modules occupied a similar footprint to a discrete solution and were expensive. More recent modules though have integrated the discrete components into smaller packages with some verti-cal stacking arrangements giving real footprint savings. SIP versions occupy little footprint but are tall

Figure 2: PoL basic schematic, single phase type

Reliable. Available. Now. www.tracopower.com

SENSORS

TBA1

HARSHENVIRONMENT

TIB80-EX

MOBILITY

IOT COMMU-NICATION

TSR1WI

TRI20

AI

TRS2

Industrial Power SolutionsDC/DC from 1–300 WattAC/DC from 2–1000 Watt§ Wide selection of mounting type & packaging

styles§ IEC/EN/UL 62368-1 compliant§ Industrial EMC immunity acc. to IEC 61000-6-2§ Special portfolio for harsh environment acc.

to ATEX, UL HazLoc§ 3-years warranty

For more information, datasheet and certifi cations please visite our website www.tracopower.com


1918


WWW.POWERSYSTEMSDESIGN.COM WWW.POWERSYSTEMSDESIGN.COM


and typically require through-hole soldering. Hybrids are also avail-able, comprising just the power components of the PoL with the controller placed separately on the motherboard, perhaps on the back side, giving some space sav-ing. Figure 3 shows components in a typical discrete implementa-tion of a PoL, occupying about 50 mm2 for 6A amp rating. Input and output filtering capacitors are not included and MOSFET switches are integrated in the controller.

Chip embedded technology – next generation PoLs With motherboards shrinking

down to smaller platforms such as in PCIe, CoM boards, SSD, Edge and AI computing, the

market demands more compact power con-version solutions. To achieve a step improve-ment, TDK has taken a ‘clean whiteboard’ ap-proach with the design of its new µPOL range of modular PoLs, for 5W to 150W systems (Figure 4). Proprietary, patented technologies have been combined to yield a dramatic im-provement in footprint and current density. The FS140x series, ini-tially available in 3A, 4A and 6A versions will be expanded in the future to 25A and on to 100A. The products avail-able now are in a tiny 3.3mm x 3.3mm pack-age just 1.5mm tall for placement in height-

constrained areas such as on the backside of the motherboard or under other components such as heatsinks and daughter boards. Power density achieved is a class-leading 1W/mm3 – 15 watts in a 3.3mm x 3.3mm space without airflow cooling. Like competing PoLs, external input and output capacitors need to be added for a full implementation but these are small due to the high switching frequency employed. From a sys-tem point of view, it is better to add the capacitors externally, as they can be dimensioned for the exact voltage and load transient response requirements of the ap-plication. Otherwise, capacitors would have to be sized for the highest possible load and high-est possible input and adjusted output voltage, adding unneces-sary cost and board space.

With the TDK µPOL in a minimal configuration, apart from input/output capacitors, no other components are required such as

Figure 5: µPOL PoL converters small form-factor FPGA power design

compensation networks, output setting resistors, bootstrap or pin decoupling capacitors.Key to the improvements achieved are TDK’s thin film inductor technology and a new technique, SESUB or Semicon-ductor Embedded in substrate, integrating the MOSFETs and control IC into the layers of a 250µm thick substrate with the inductor placed on top. There are no wire bonds, making the devices very robust, and the parts are light, therefore suitable for drone, machine vision and other small form-factor applica-tions. Thermal characteristics are controlled so that heat from the IC and inductor is effectively led through the substrate termina-tions to the motherboard. This is so effective that at full 6A output, the module temperature rise is just 41°C (12V input and 1.8V out-put).

System benefits with the µPOLThe tiny footprint of the µPOL en-ables the smallest of spaces to be utilised for voltage regulation close to target devices. Out-put voltage is select-able from 0.4V to 5V and input voltage range is 4.5V to 16V with a single supply and 2.5 to 16V with an extra auxiliary supply providing a bias voltage. Appli-

cations are scalable, as the 3A to 6A parts have the same footprint.

Integrated monitoring and control functionality in the µPOL also add significant system benefits: for example, a ‘power good’ output and ‘enable’ input can be utilised for ‘ratiometric’ or ‘simultaneous’ power sequencing of multiple µPOLs. Protection is comprehen-sive with output over-voltage, in-put under-voltage and output over-current or short circuit monitored, all with selectable thresholds, with ‘hiccup’ mode shutdown. The de-vices operate over the range -40°C to +125°C with programmable ther-mal shutdown and auto-recovery.

The built-in I2C interface supports Fast and Fast Plus modes and en-ables a wide range of settings via PMBus commands:

• Settable output voltage, 0.4V to 5V with 5mV resolution

(fixed input versions initially available – variant dependent)

• Output voltage trim, (+/- 0.5% initial accuracy)

• Optional soft start and stop rates 0.5mV/µs or 1mV/µs

• Adjustable PGood Threshold: 85%, 90% and 95% of Vout

• Adjustable OVP threshold: 105%, 110%, 115% and 120% of Vout

• Adjustable OVP scheme: Latch, Unlatch

• Adjustable OCP threshold: up to 8A with 250mA resolution

• Adjustable OTP threshold: 75oC, 85oC, 125oC and 145oC

• Selectable ‘continuous’ (CCM) or ‘discontinuous’ (DCM) operating modes for EMI and efficiency optimization

With a wide range of devices, op-tions and settings, system design-ers can ‘tune’ applications using the µPOL for minimum system cost and maximum functionality.

Application exampleFigure 6 shows the TDK µPOL con-figured for Vin = 12V, Vout = 1.8V and Iout = 6A with ripple voltage requirement less than +/-1% pp Vout and a maximum Vout excur-sion of +/-3% with a 3A load step. The µPOL part is FS1406 and 2 x 22µF capacitors are initially chosen for both Cin and Cout. Switching frequency Fsw is 2MHz.

Input capacitor selection depends on ripple current handling IRMS. In a buck converter operating at duty cycle D, this is given by:

Figure 6: The TDK FS1406 µPOL in a typical application

Figure 3: A discrete implementation of a 6A PoL converter with necessary peripheral components excluding input/output capacitors

Figure 4: The µPOL from TDK with a footprint of just 11mm2 and 1.5mm height

20



Figure 7: Ripple, application example, 6A, 1.8V output

Figure 8: Transient load response, +/-1.7% excursion with a 50% load step

Figure 9: An example of TDK Power Strip Design Board, 8 output µPOLTM units with sequencer board.

Here, Iout is 6A and D = Vout/Vin = 0.15, giving IRMS = 2.14A which is easily handled by two 22µF, 16V capacitors type TDK C3216X5R1C-226M160AB.

The minimum number of output capacitors required, NMIN to meet the target ripple Δ Vout is given by:

TDK MLCC capacitors C2012X5R-0J226K125AB have an ESR of 3 milliohms, an ESL of 0.44nH and an effective capacitance of 12µF at 1.8V. Solving, NMIN is 1.27 so 2 ca-pacitors as assumed is adequate.

The minimum number of capaci-tors to meet the load transient re-

sponse ΔVoutmax with load step ΔIout is given by:

Solving again with 12µF, NMIN is 0.4 so two TDK ca-pacitors C2012X5R-0J226K125AB is a good solution.

The resultant ripple and load transient response plots are given in Figures 7 and 8 showing just 15.19mV pp ripple or 0.8%, and about 30mV excursion or +/-1.7% with load steps.

At higher output voltages, the

minimum number of 22µF output capacitors decreases for the same ripple and transient load specifica-tion and above 2.25V output, just one is required.

TDK µPOL - includes comprehensive supportTo support design-in of the

µPOLT, TDK provides a suite of tools including general purpose evaluation boards with interfaces

to access the I2C bus, with an easy-to-use GUI to set up the

parts. Additionally, ‘Power Strip Design Boards’ are also available with up to eight µPOL units fitted to provide all the rails, includ-ing sequencing boards, for Xilinx Spartan 7, Artix 7, Zynq 7, Zynq UltraScale+, Zynq UltraScale+ RFSoC, Versal ACAP (Versal Prime Series) and other FPGAs from ven-dors such as Altera, NXP (i.Mx8, LS series), Marvell/Cavium and Microsemi. For quick implemen-tation, schematics are provided along with PCB ‘layout snippets’ for ORCAD/Allegro, Altium and Mentor PADs+/Xpedition CAD programs. ‘Ready to go’ power subsystem solutions for vendor FPGAs are available from TDK and, as a Xilinx ‘Alliance Program Certi-fied member’, TDK has generated complete reference designs for FPGAs such as the Zynq7 series.

Avnet Abacuswww.avnet.com/wps/portal/aba-cus/manufacturers/m/tdk/prod-ucts/fs1406-µpol-buck-regulator

Special Report:Industrial Applications

Inside:

Switches: Critical to Modern Industrial Sensors

FPGA Power System Management...

Smart Connected Sensors Drive Innovation on the Factory Floor...

The Benefits of Hybrid Regulated Ratio Topologies for Intermediate Bus

Converters...

Making Cables Safer...

Supercapacitors Deliver Reliable Power for Warehouse Automation...

22

25

29

32

38

42

2322 WWW.POWERSYSTEMSDESIGN.COM WWW.POWERSYSTEMSDESIGN.COM

SPECIAL REPORT : INDUSTRIAL APPLICATIONS POWER SYSTEMS DESIGN 2020APRIL

Switches: Critical to Modern Industrial Sensors

By: Craig Thomson, Director of Global Portfolio Management and Markets at C&K

Robotics, artificial intelligence, machine learning and industrial IoT applications are leading the new age of automation

Machines can now outperform their human counterparts in

activities that range from parts assembly to more cognitive tasks including self-learning algorithms for data mining and pattern recognition. At the center of this new industrial revolution are sensor switches that can simplify both factory and process automation. Coupling sensor technology with mission-specific switches is the next logical step in this automation evolution. By giving sensors the ability to control the operation of the system, it removes the need for human monitoring of process flows, freeing-up personnel for more crucial tasks.

The Modern Industrial SettingA report from Market Watch recently suggested that factory automation has managed to reduce both the production time and cost of industrial goods. The publication notes the segment, “…is expected to register a CAGR of over 8.4% during the forecast period 2019 – 2024.”

In this competitive global environment, companies need to focus on their ability to innovate. With the popularity of emerging technology accompanying Industry 4.0, the need for keeping up with innovation is now more critical than ever.

Sensors and switches are critical components in any modern automation process. Today sensor designers and manufacturers bringing the two components together to meet specific challenges. Critical safety switches enable personnel to shut down lines in the event of a failure of central systems, preventing damage to equipment and loss of life. Process control switches, on the other hand, can help to control which subsystems are operating and which are switched off.

According to the new market

research report "Process Automation & Instrumentation Market by Instrument” from Markets and Markets,” the process automation and instrumentation market will be worth over $95 billion by 2024. The reason automation is being adopted in such a broad fashion by industry is because it solves two fundamental problems process plants face:

1. The rising cost of labor makes automation an attractive option for businesses trying to control staff costs.

2. Safety in plants is of the

utmost concern since companies face massive backlash in the event of a catastrophic failure of safety protocols.

Sensors will play an integral role in facilitating this adoption.

Typical Automation SensorsAutomation sensors are used across industries and in varied environments for factory automation and process automation. While each industry has specific needs, the array of sensors can include:

• Proximity sensors: A proximity sensor can be either inductive or capacitive, depending on the type of material that the sensor is tasked with detecting

• Position sensors: Typically using photo-electric technology, these sensors are used to detect the position of various components in the assembly process, and are helpful in a variety of industrial applications, including the automotive sector

• Pressure sensors: Using piezoresistive or equivalent technology, these sensors monitor pressure in tanks and pipes and are a critical safety component in the processing industry

• Flow sensors: Automated flow switches monitor

the flow rate of a fluid and determine whether to switch a subsystem on or off as required

Sensors Rely on Good SwitchesWhere sensors are the nerves in automation, switches of varying sizes, and form factors facilitate

human interaction with those sensors. Switches are coupled with almost every type of sensor application. For instance, tact switches like C&K’s KSS and KSC3 can be used with a display screen or an LED-based menu for sensor setup and calibration.

Figure 1: Rated to IP67, the KSC3 series sealed tact switch offers positive adaptable tactile feeling

24 WWW.POWERSYSTEMSDESIGN.COM

SPECIAL REPORT : INDUSTRIAL APPLICATIONS

Figure 2: The KSS series subminiature SMT side actuated switch features a large flat top, a low profile of 1.7 mm, and can withstand high shear forces for products that are exposed to harsh environmental conditions

Figure 3: Built to meet the specific needs of engineers creating the next generation of devices, the KXT3 series ultra-low profile top actuated tactile switch is reliable and long-lasting

Miniature top actuated switches, like C&K’s KMT0 and KXT3 line, work well in applications requiring a small footprint or an ability to withstand harsh environmental factors. They are also ideal as “teach” buttons for sensor calibration and provide tactile and audible responses to the user.

Sealed tactile switches, such as the KSC2 and KSC3 from C&K, are designed for applications in consumer, automotive, and industrial

automation applications. While there are typically no switches on capacitive or inductive proximity sensors, DIP and tact switches are frequently used in the signal conditioners for these types of sensors.

For applications that require logic configuration or node addressing, standard and rotary style DIP switches are often used. Both types of switches are designed for low level logic switching and can be invaluable in IoT implementation. Because of their low profile and minimal power consumption, a DIP switch can be incorporated into most IoT applications.

Of course, there are also switches that complement factory automation to control displays on sensors, controllers, and other automation systems. Switch manufacturers have designed highly responsive microminiature SMT top actuated tact switches with excellent tactile feel to meet these demanding application requirements.

For heavy industrial applications such as pressure transmitters used in tanks and piping systems, both tact and dip switches are specially made with custom tactile feedback and response to meet environmental requirements.

The Overlooked Necessity of SwitchesProcess plants that move towards increased automation need to have high-quality, dependable switches for the long-term success of their endeavor. While the switch is usually overlooked as a component, choosing the wrong switch for a sensor application can increase maintenance issues, user error, and reliability of the sensor. The right switch coupled with a sensor application can yield a more efficient control automation solution with higher accuracy and throughput.

C&Kwww.ckswitches.com

FPGA Power System Management

By: Pinkesh Sachdev, Product Applications Engineer, Analog Devices

Field-programmable gate arrays (FPGAs) trace their origins back to the 1980s, evolving from programmable logic devices (PLDs)

Since the 1980s, FPGA resources, speed, and efficiency have improved rapidly,

making FPGAs the go-to solution for a wide variety of computing and processing applications, especially when production volume doesn’t justify application-specific integrated circuit (ASIC) development costs. FPGAs have advanced to such an extent that they’ve also found homes in large-scale deployments. For instance, after successfully speeding up the Bing search engine with FPGAs in a 2013 pilot program, Microsoft® expanded FPGA-equipped server usage to its cloud data centers.

FPGA Power System RequirementsFPGAs require a few different low voltage supply rails, each with its own voltage and current specification, to power their internal core logic, I/O circuits, auxiliary logic, transceivers, and memory. These rails may need to turn on and turn off in a specific sequence to avoid damaging the FPGA. Point-of-load (POL) regulators step down the board’s higher input supply voltage to the multiple lower rail voltages required by the

FPGA. Switching regulators are used as POL regulators when power conversion efficiency is paramount, whereas linear regulators—for example, low dropout (LDO) regulators—are employed for noise-sensitive circuits such as PLLs and transceivers.

Typical board input voltages are 5 V, 12 V, 24 V, and 48 V, while FPGA rail voltages range from below 1 V to around 3 V. For high input voltages (12 V, 24 V, 48 V), an extra step-down may be needed to generate an intermediate voltage bus that feeds the POL regulators (see Figure 1). Among the FPGA rails, the core supply requires the lowest voltage (around or below 1 V) and highest accuracy (±3% or better), with current levels in the tens of amperes depending

on FPGA resource utilization. To prevent logic errors, the supply variation needs to be limited to tens of millivolts, as dictated by the FPGA rail tolerance specification, not just under dc conditions but also during FPGA current transients. The worse the power supply’s dc accuracy, the more bypass capacitance is needed to maintain an acceptable supply voltage under transient conditions. For example, assume a ±3% core voltage tolerance specification. Using a ±1% accurate dc supply leaves a good ±2% allowance for transients. On the other hand, a less accurate ±2% dc supply leaves less room (±1%) for transients, requiring more bypass capacitance than the previous case.

Tweaking or trimming of the FPGA supply voltage level

Figure 1: One possible FPGA power tree: a high voltage input supply (for example, 12 V, 24 V, or 48 V) is stepped down to an intermediate voltage bus feeding the POL regulators that power the FPGA





is a power system management solution that is simple, scalable, and flexible.

Digital Power System ManagementDigital power system management (DPSM) devices deal with the complex power systems found on digital processing boards. They are available with and without integrated dc-to-dc conversion to either replace POL regulators or work with existing POL regulators. Power system managers—that is, without dc-to-dc conversion—add digital monitoring and control to any existing analog power system, whether made up of switchers or LDO regulators. For example, a single device such as the LTC2980 from Analog Devices trims, margins, monitors, sequences, supervises, fault logs, and fault manages 16 POL regulators. Differing channel-count devices (2, 4, 8, or 16

channels) can be mixed and matched to manage hundreds of rails. Similarly, the 2-channel LTC2972 provides a simple introductory solution for monitoring and controlling the two most critical rails in such a power system; for example, the FPGA core and auxiliary rails.

2-Channel Power System ManagerA 2-channel power system manager adds comprehensive software-based monitoring, control, and black box fault recording to the power systems of FPGA, ASIC, and processor boards, accelerating time to market, enhancing system reliability, and optimizing board energy consumption (Figure 4). In a device like the LTC2972, POL supply output voltages are trimmed, margined, and monitored using a best-in-class 16-bit ADC with 0.25% total unadjusted error (TUE), improving board yields and long-

term performance. The ability to tighten POL output voltage to ±0.25% accuracy leaves plenty of room for it to move during load transients (±2.75% for a ±3% FPGA rail specification), significantly reducing needed bypass capacitance and freeing up board space. Supply output currents are measured using a sense resistor, inductor DCR, or the IMON output of a power supply. The voltage and current measurements are multiplied internally to provide a convenient POL power output reading.

Supply sequencing, supervision, and EEPROM fault-logging are built into the LTC2972. Sequencing is accomplished with time delays written to an internal register or with cascaded power-good signals. Dedicated fast comparators signal faults when POL input voltage, output voltage, and temperature stray outside digitally settable low and high thresholds. Faults trigger EEPROM black box recording, simplifying failure analysis while providing insight into future system improvements. A first fault command provides additional insight into causes of system failure. Faults can be flexibly propagated to other supplies or other DPSM devices.

A power system manager features voltage, current, power, and energy monitoring of the intermediate bus input to POL converters. Monitoring circuit board power and energy use is a prerequisite for managing, optimizing, and

Figure 4: A 2-channel power system manager like the LTC2972 provides intermediate bus energy monitoring and POL output power monitoring

around the default setpoint is required for last-minute design changes, design reuse in another application, board margin testing, and dynamically optimizing system power consumption during development or field operation. Soldering in different resistors in the supply’s feedback network isn’t the quickest or most feasible solution for such situations. One method to achieve voltage trimming is with a digital-to-analog converter (DAC) driving the feedback network of a voltage regulator (see Figure 2). Software code needs to be written for the trim routine to obtain supply voltage measurement data from an analog-to-digital converter (ADC), to compute the correct

DAC code, and then slowly adjust the DAC output to the computed code for smoothly ramping the supply voltage, without glitches or overshoots, to the target level. This trim routine needs to be repeated over time to ensure that the supply doesn’t walk away from the target voltage due to components drifting with time or temperature.

Monitoring FPGA supply voltages, currents, and faults is essential to understanding system health and power consumption under different scenarios because the FPGA is the brain of the electronic system. Such an understanding, coupled with trimming capability,

avoids designing supplies for the worst case, saving cost and power. Moreover, an upcoming system malfunction could show up as an abnormal trend in FPGA power consumption, alerting the host controller or service staff before the board or system goes down. Voltage monitoring requires ADCs, whereas current monitoring also requires level-shift circuits to translate the high-side current sense voltage to a ground-referenced voltage; for example, with a transconductance amplifier, as shown in Figure 3.

One’s head may be spinning after reading this long laundry list of requirements although we have not discussed fault management. What should happen when a POL output goes undervoltage or overvoltage—that is, outside the valid voltage window? Should only the faulting supply be turned off, or should other supplies be turned off too? How does one debug a fault that has shut down the board?

As one can see, managing an FPGA’s power system can become complicated very quickly, distracting from the essential FPGA application. Remember that the FPGA’s power tree is just a portion of the overall power system on a digital processing board. Most of the above requirements also apply to other digital devices such as ASICs, DSPs, GPUs, SoCs, and microprocessors. What is needed

Figure 2: POL supply output voltage trimming to target voltage with DAC and ADC

Figure 3: One possible discrete circuit for monitoring POL supply output voltage, current, and power

Figure 5: Development environments like LTpowerPlay for DPSM devices eliminate the need to code for autonomous operation.

reducing their consumption in order to lower server and data center cooling and utility costs. The power system manager relieves the host of burdensome polling and computation by conveniently providing the input energy, reported in joules, and the elapsed time through a PMBus interface, the industry standard for communicating with power management and conversion devices. When combined with its digital measurements of POL output voltages, currents, and power, the use of a power system manager enables long-term monitoring of a power system’s conversion efficiency.

Programmable power-good, or general-purpose input/output (GPIO), pins are available with each channel. Multiple power system managers can be connected together to coordinate sequencing and fault management

of more than two rails. PMBus-compatible commands over an I2C/SMBus interface are used for flexible programming and data readback of the power system. Configuration is managed through the development environment, such as LTpowerPlay which supports all of Analog Devices’ DPSM products (see Figure 5). Once the internal EEPROM is programmed with the needed application-specific configuration, no further software coding is required for autonomous operation.

FPGAs are spreading to all kinds of electronic systems, even taking over the job of ASICs, but they come surrounded by a complicated power system. Digital System Power Management devices can help manage this complexity.

Analog Deviceswww.analog.com/FPGA

www.powersystemsdesign.com

THE BRIGHT SOURCEFOR ALL THINGS POWER.

Smart Connected Sensors Drive Innovation on the Factory Floor

By: Sergey Komarov, Business Development Engineer, TT Electronics

Optimising processes with data in context brings new performance to Industry 4.0

Connected devices and smart equipment are streamlining manufacturing and

supply chain processes on a global scale. Smart sensors are at the heart of this movement, driving change by extending the inherent capabilities of legacy sensors. Today, sensor technology that merely collected, stored and monitored data can now optimise processes by capitalizing on data in context.

This is no easy feat. Performance environments for manufacturing and assembly machinery are often dirty and characterised by fluctuating and extreme temperature conditions. Ambient lighting might include full daylight as well as periods of dark and bright lighting. Media often has low-contrast reflectivity, even as packaging, printing, and labelling equipment commonly require robust presence detection. Reliable presence detection is also key to logistics and accurate routing. At the same time, system engineers need sensors with proposition value that comes with smaller circuit designs that are highly

integrated yet energy efficient, simple to implement, and easily optimisable (programmable).

Versatile performance adds automation valueSmart optical sensors offer ideal performance in meeting these diverse industrial challenges. Their extreme versatility comes from the ability to look at objects (such as products on a manufacturing line), through media (such as liquids flowing through tubes), or at reflections off things (such as surfaces that have been processed in some way). For analysis to be meaningful, measurements must be consistently accurate and repeatable – potentially very challenging given the environments and materials in use. Making sense of the resulting signals calls for a sensor with programmable sensitivity levels and output that effectively manages the varying levels of reflectance and low contrast settings.

Sensor signals are often quite ‘delicate’. As small currents or voltages, they are measured in electrically noisy conditions,

generated by sensing devices that change performance over their operational lifetime based on environmental factors such as ambient light or temperature. For example, the sensor’s LED performance is subject to ageing since light output in LEDs degrades gradually over time. An adaptable sensor reduces this performance challenge by simply adjusting – the sensor automatically recognises and re-calibrates to this change. As a result, its readings remain consistent throughout the life of its deployment, adding critical value in complex manufacturing settings where downtime from sensor failure is costly and could even shut down operations.

Powering the factory floor with optical sensors On the manufacturing floor, connected devices are essentially creating networks of highly instrumented devices – used to implement one or more sensors to recognise changes in diverse, real world environments and turn them into measurable, actionable electrical signals. Readings must be accurate, facilitated by sensors that are highly integrated, small,





applications. This optical sensor comprises a fully integrated analog front end, on chip processing and digital interface in a surface mount package measuring just 4.0mm by 2.2mm by 1.5mm – a space savings of as much as 80% in circuit complexity. Superior integration solves many factory challenges, transforming a standalone sensor into a low-power sensing module that contains all its supporting circuitry and has programmable sensitivity and thresholds. Inherent flexibility enables these kinds of devices to serve many markets in a more cost-effective manner than their discrete alternatives. Going even further, TT recently made available the OPB9001 series for easy implementation; these sensors integrate the OPB9000 sensor – with all the required passive components on a single

PCB – for a variety of operating supply voltage ranges from 4.5V to 30V.

With market-leading ambient light immunity of 25klux, this sensor can be used where others cannot. It also integrates a variety of features not found in other devices, such as self-calibration and Automatic Gain Control, temperature compensation circuitry, programmable contrast sensitivity, and an industry standard communication interface. With a response time of just 6µs, it is extremely fast – without the use of microcontrollers and outputs which results in substantial cost savings. While mass produced, these sensors can be programmed to the specific application, offering additional cost savings and empowering first to market

customization capabilities.

Smart factories require smart technologyIndustry 4.0 thrives on flexible technology, demonstrated by the industrial sewing application featured above. Thread tension, needle force, even measurements such as thread usage are all elements that can be managed with smart, integrated sensors. Ultimately, sensors manage factors that keep machines running safely and at optimal performance.

These tiny devices play an outsized role on the connected factory floor, empowering manufacturers to tap into real-time performance data, in turn driving continuous improvement in a manufacturing landscape that remains highly competitive. And because industrial settings are so diverse, performance critical and potentially dangerous to human workers, smart sensors are credited with delivering value while reducing risk.

By measuring and sensing a range of conditions while consuming minimal power, optical devices are ideal for industrial operations. These highly integrated devices are small and rugged – critical factors in improving equipment portability and handling even more challenging factory environments.

TT Electronicswww.ttelectronics.com

Figure 2: The Photologic V OPB9000 reflective optical sensor from TT Electronics is designed for diverse applications and includes a fully integrated analog front end, on-chip processing, and a digital interface in a surface-mount package measuring just 4.0 × 2.2 × 1.5 mm

robust, stable over the long term, and draw little power. It is this unique combination of features and performance that is critical in keeping industrial applications operating consistently, safely and with minimal downtime.

Industrial sewing is a demonstration of sensors in action, powering sophisticated systems that are worlds apart from pedal-controlled, mechanical sewing machines used by consumers. Advanced sewing systems are electrical rather than mechanical and may operate non-stop on the factory floor, capably handling a range of stitches on a variety of fabrics with diverse textures and general characteristics. Like any type of

electrical system, safety must be engineered into the product: in our example, sensors are tasked with detecting the presence of fabric and activating the machine only when fabric is centred over the throat plate of the device. This operation assures the needle pierces the fabric at the appropriate speed and depth, and reduces thread waste by powering on only when the fabric is correctly positioned.

Adaptable sensors adjust the machine’s speed and intensity to accommodate lightweight or more transparent fabrics, as they are programmed to recognise different types of fabric. Because room temperature on the factory floor can sway dramatically

between hot and cold, adaptable sensors have built-in temperature compensation circuits which sense their environment and respond by automatically re-calibrating and adjusting their output based on the status of their environment. This same automatic adjustment applies to ambient or shifting light conditions common in industrial settings. An optimised adaptable sensor can continue to perform flawlessly by detecting the light coming only from its own LED, effectively screening and filtering out the ambient light from its working environment.

An example of such a device is the TT Electronics Photologic V OPB9000 reflective optical sensor, expressly designed for industrial

Figure 1: Sensors must offer versatility for broad applicability



The Benefits of Hybrid Regulated Ratio Topologies for Intermediate Bus Converters

By: Oscar Persson, Specialist R&D Engineering, Flex Power Modules

Hybrid regulated ratio (HRR) technology is a compelling alternative to conventional fixed-ratio or fully regulated conversion

When designing the distributed power architecture for applications

such as data centers, an isolated intermediate bus converter (IBC) might be used to take a nominal 48V or 54V input and generate a 10V to 12V bus for one or more downstream point-of-load (PoL) converters. In the past, there have been broadly two approaches – a simple fixed-ratio converter with unregulated output, or one with full regulation with fixed output.

Now, hybrid regulated ratio (HRR) converters have been developed to combine the advantages of both, switching seamlessly between each mode according to the input voltage. To fully understand the value of HRR, it is necessary to compare the operation of HRR, fixed-ratio and fixed-output converters and consider each in relation to typical operating conditions.

Background and operating principles

Recall that pulse width modulation (PWM) of a bridge-type switching power converter controls output voltage – longer switch on-times increase the average voltage passed to the output while shorter on-times reduce the average with an output filter smoothing the waveform to DC. Sensing output voltage, PWM can be used to compensate for varying input voltage, with the transformer turns ratio set to give close to 100% duty cycle at minimum input. Higher loads also increase duty cycle incrementally as the converter compensates for increased voltage drops at higher current. At near 100% duty cycle, little energy needs to be stored during the off-time in the output inductor, and input and output switching currents are close to continuous, with their average value close to the RMS value. In contrast, at low duty cycles, significant energy must be stored and released each switching cycle, and current waveforms have high ratios of peak to average with consequently high RMS values, with both effects increasing losses.

Highest efficiency of regulated converters is therefore typically seen at minimum input and high load. At higher inputs and light loads, producing low duty cycles, efficiency is compromised. Regulated converters can however be set to give a fixed, optimum output voltage for downstream PoLs, recovering system efficiency somewhat. Input range can also be wide, easily covering the ETSI standard 36V to 72V.

Fixed-ratio converters operate at near 100% duty cycle for all conditions, needing only a very small output storage choke, and can be efficient across a wide input range with their lower input and output RMS currents. Output capacitors can also be smaller for the same ripple voltage, as ripple current is always low. Input range normally has to be restricted, though – with a 5:1 transformer ratio for example, the output would vary from about 7V to 14.4V for 36V to 72V input. This would constrain PoLs to be wide-input types, which are less

efficient, especially at 7V input where currents are higher for the same PoL output load power. For this reason, fixed-ratio converters are often specified for a narrower input range, typically 45V to 54V, limiting their usefulness. A further disadvantage is that voltage surges on fixed-ratio converter inputs pass straight to the output, causing potential stress on PoLs and EMI.

HRR is the best of both worldsThe HRR technique offers the best of both approaches; the converter operates at near 100% duty cycle Figure 1: HRR operation is similar to fixed ratio at lower Vin, but limits Vout

at higher values of Vin

Figure 2: HRR converter maximum output power occurs when Vin is close to the nominal value of 54V

Figure 3: HRR converter efficiency peaks at nominal input voltage

at low input voltages, but moves seamlessly to PWM control above a set point, limiting the output voltage. The effect is shown in Figure 1, comparing the output voltage from a typical 5:1 fixed-ratio converter with a HRR version such as certain variants of the PKU-D series, limiting at 12V output as input voltage varies from 35V to 75V. The fixed-ratio converter is forced to shut down above 65V to limit the output voltage to a safe value. The HRR device with a 4:1 ratio transformer is dimensioned to remain in 100% duty cycle mode up to the nominal input operating voltage, 54V in this case, resulting in a higher output between 36V and about 60V input compared with the fixed-ratio converter. In fact, it can take short-term overvoltages up to 80V without losing output control.

Examining HRR operation under various conditions can reveal how the converters are used to best advantage, quantifying the benefits in terms of efficiency and system design.

HRR and fully regulated converter comparisonFigure 2 and Figure 3 illustrate the HRR converter’s power capability and efficiency. Figure 2 shows that for fixed maximum output current, HRR can be designed for the highest power at the nominal Vin of 54V with some derating at high voltages



due to switching losses reducing efficiency.Figure 2 here

Figure 2: HRR converter maximum output power occurs when Vin is close to the nominal value of 54V

The efficiency curves in Figure 3 show that a HRR converter is also increasingly efficient as Vin rises to the nominal 54V, because of reducing resistive and other voltage drops, as the current reduces with increasing output voltage, at constant load power. This is especially true at high load. Note that although the efficiency and maximum output power are both reduced at the lower values of Vin, the performance remains impressive. Efficiency of 95.1% when delivering 200W from 40Vin is good performance for an intermediate bus converter.

Comparing the performance curves across the range of Vin values with those for a fully regulated fixed-output 12V converter, in this case a Flex Power Modules PKU4213D,

highlights a key advantage of the HRR converter: Figure 4 shows the set of efficiency curves for the Flex Power Modules PKU4213D, indicating that the best performance

is at low Vin where duty cycle is higher. However, intermediate bus converters rarely operate with such low values. In practice, Vin is typically around the nominal value most of the time, and this is where the HRR converter achieves best power capability and efficiency. This delivers major advantages over the fully regulated converter for end users, including lower utility bills, lower operating temperatures, reduced demand for cooling, better reliability and longer equipment operating lifetime.

Note that even at lower values of Vin, HRR outperforms the conventional technology. Efficiency is 1.2% higher at 150W and 40V input. At 200W and 54V input, this improvement increases to 2.2%.

Fixed-ratio converter comparisonHigh efficiency is the main attraction of using a fixed-ratio converter with duty cycle always near 100%, but this only applies for a limited range of input voltages. However, the HRR

technique compares very favorably and, in some cases, can be superior.

Covering an input range up to 60V will likely require a 5:1 fixed-ratio converter. In this case Vout will be approximately 10.4V at 54V input, which provides a good combination of efficiency and dynamic performance. A comparable 4:1 HRR converter could be selected with a limiting Vout of 12V or 10.4V (for example the PKU4217D). The 12V option transitions to ratio mode at around 50V, delivering very similar performance to the fixed-ratio type when the input is nominal at 54V. The 10.4V option transitions at about 43V, giving a slight performance penalty for the converter at 54V where the duty cycle has decreased significantly. However, the lower voltage to following PoLs is likely to make them more efficient, so the overall effect may be minimal.

If Vin dips, however, so does the performance of the fixed-ratio converter. At 40V input, a 5:1 fixed ratio converter produces about 7.7V output. This forces the converter to operate with 35% higher output current than at nominal Vin/Vout for the same load power. The HRR module reduces to only about 9V so the current increase for fixed-load power is less, yielding a significant saving in resistive distribution losses.

In this scenario, while the 10.4V

HRR converter might have a small disadvantage in maximum power at nominal Vin, it has a flatter Vout profile across the input voltage range and a more consistent output power curve. Put simply, this means that it can deliver more power under conditions where it is most needed.

Handling input transientsWhile the HRR technique has been shown to deliver valuable gains in efficiency and power under the most common operating conditions, its response to steep positive-going input voltage transients can be made superior to fixed-ratio converters, making them particularly attractive in the

telecom industry.

Figure 5 compares the responses of three different DC-DC converters to a steep input transient from 48V to 54V.

The upper trace shows the response of a PKU4213D 17A converter with fully regulated fixed-output voltage of 12V powering an 8.5A load. There is minimal change in Vout, as would be expected.

The second trace is from a 12V, 25A HRR converter with a 12.5A load. The smoothly ramping Vout takes approximately 4 milliseconds to adapt between output voltages, in this case transitioning between “ratio” mode and constant output voltage mode.

The lower trace, for a 9.6V 42A fixed-ratio 5:1 converter with 12.5A load, shows Vout following Vin with a step-change in voltage.

A following PoL converter may pass the step-change in input voltage from a ratio converter to the PoL load as a transient, causing functional problems or even damage. The HRR technique with its active control can momentarily reduce the duty cycle and slow the IBC output voltage change before reverting to 100%, thus protecting the PoL.

The step voltage that can be passed by a fixed-ratio converter also causes high transient current

Figure 4: Efficiency of a fully regulated IBC is typically best at low input voltages

Figure 5: Fully regulated, fixed-ratio and HRR converter responses to a 48V to 54V Vin transient

Figure 6: Output capacitor current response to a Vin transient for fixed-ratio, fully regulated and HRR converters



into the IBC output capacitors. Figure 6 shows example current in two 470µF output capacitors during the Vin transient from 48V to 54V for the 5:1 fixed ratio converter considered before.

For the fixed-ratio converter, the charge current to the capacitors reaches 36.1A due to the very sharp Vout change. The converter is overloaded (12.5A + 36.1A

> 42A) and output droops significantly. Doing the same test with a step from 40V to 54V causes the converter to enter overcurrent protection.

With the HRR converter, the capacitor charging current is just 0.7A, which compares well with the charge current of 0.3A for the fully regulated fixed-output converter. This is one of the

important benefits of HRR. The PKU-D HRR products from Flex Power Modules actually have a constant current characteristic on overload, with shutdown and restart if the overload persists.

Negative transients can be troublesome in ratio convertersMost modern converters use the output synchronous rectification (SR) technique to replace discrete diodes for best efficiency. The technique does have a disadvantage, though; the MOSFETs configured as rectifiers can sink as well as source current. If a fixed-ratio converter experiences a negative-going input transient, the output immediately tries to drop, and because the 100% duty cycle switching is maintained, current can flow backward from the charged output capacitors, reflecting to a high reverse current on the input. This can also affect a HRR converter, but a feature of the Flex Power Modules HRR parts is reverse current detection that, in the worst case, sends the converter into standby mode until the output voltage falls to the correct level. The load is unaffected as the output is still, by definition, held up by the output capacitor in this situation.

Input ripple rejection is better with HRRFor fixed-ratio converters, any low-frequency ripple voltage on the input is passed straight through to the output in

proportion to the ratio, with even the possibility of amplification due to “peaking” in input and output filters at higher frequencies. With the HRR technique, ripple is similarly attenuated by the conversion ratio at low frequencies but, because the duty cycle is never quite 100%, some “feed-forward” PWM can be applied to attenuate ripple. Figure 7 shows the resulting ripple attenuation of a HRR converter with the expected 4:1 attenuation up to 1kHz increasing to around -35dB at 10kHz.

Compare this with the attenuation plot from a fixed-ratio converter in Figure 8 with the 5:1 (-14dB) attenuation up to 1kHz but just -10dB at 10kHz.

The peaking at higher frequencies up to 100kHz due to input and output filter resonance in an

Figure 7: HRR converter – input ripple voltage attenuation

Figure 8: Fixed ratio converter – input ripple voltage attenuation

Figure 9: Input and output filters can produce amplification of input ripple voltage

Figure 10: Flex Power Modules PKU4217D

example fixed-ratio converter is shown in Figure 9, resulting in 350mV output ripple for 200mV input (+5dB) at around 20kHz.

ConclusionNeither a fixed-ratio nor a fully regulated fixed-output converter can provide an entirely satisfactory way of generating an isolated intermediate bus to power non-isolated PoL converters. A fixed-ratio converter, although efficient under certain circumstances, cannot handle the full range of input-voltage variation required by ETSI standards and does not handle input transients well. A fixed-output converter is relatively

inefficient around the nominal input voltage.

The Flex Power Modules proprietary hybrid regulated ratio technique enables greater efficiency and power delivery under the most common operating conditions, permits a wide input voltage range, and provides superior response to transients with low output ripple. These advantages ensure strong performance in telecom, industrial, laboratory and other applications where distributed power architectures

with isolated IBCs are used. Converters from Flex Power Modules incorporating HRR technology are available in power ratings from 260W to 1300W, with signific

0 3 5 ) . & 3€¦ · Engineering, Flex Power Modules Making Cables Safer By Adam Franczak, TME...

Documents

Transcript of 0 3 5 ) . & 3€¦ · Engineering, Flex Power Modules Making Cables Safer By Adam Franczak, TME...