Where fixed-ratio converters fit in high-power delivery systems Page 20
The try-before-you-buy route to energy efficient power design Page 38
FEBRUARY 2020
Power Electronics Handbook
200109_AutoKuku1_EEW_US.indd 1
1/6/20 2:07 PM
™
hIgh
PERfORmaNCE
Battery Clips, Contacts & Holders
• Leading-Edge Battery Clips, Contacts & Holders to accommodate all major manufacturers’ batteries • Products available for Coin/Button cell, Cylindrical batteries, 9 Volt, Alkaline, Lithium, Lithium-Ion and Rechargeables • Manufactured in a variety of materials and mounting styles including On-Board and Off-Board applications • Applications include Surface Mount (SMT), Thru-Hole (THM), Solderless locking PC Connectors and Molded Cavity Contacts • Ideal for all industries and designs It’s what’s on the InsIde that counts ® E L E C T R O N I C S
C O R P.
View our Dynamic Catalog M70 at www.keyelco.com (516) 328-7500
(800) 221-5510
Request a copy of our Product Design Guide M70
EE-DWTHiNK Fist+BCCH 2-20rFinal.indd 1
1/10/20 11:31 AM
POWER ELECTRONICS HANDBOOK
Why we don’t need energy eff iciency standards for lighting HEADLINES WERE MADE RECENTLY when two national associations dropped a lawsuit attempting to squash light bulb efficiency standards in California. The California standards supposedly save Californians billions of dollars on their energy bills and avoid millions of tons of carbonwarming pollution. The suit was brought by NEMA (National Electrical Manufacturers Association) and the American Lighting Association. At issue were lighting regulations which had been expanded to cover specialty light bulbs like those used in bathroom vanities and recessed lighting, as well as candle-shaped lights. The DOE recently eliminated these bulbs from national lighting efficiency standards. Green advocacy groups crowed about this legal development. Said the director of the Center for Energy Efficiency Standards at the Natural Resources Defense Council, “The lighting industry finally came to its senses and discontinued its desperate efforts to block California’s common-sense light bulb efficiency standards, which are poised to save consumers billions of dollars on their utility bills.” Hold on there, Kemosabe. The thing to note about specialty light bulbs is that they are frequently in places where they’re not turned on all that often. As is in any energy consuming application, energy costs are only one part of the equation. The initial cost of the bulb also factors in. If an energy efficient bulb costs sufficiently more than a less-efficient version, buyers may be better off spending more on energy, particularly if the the bulb isn’t used much. This reality hit home for me because of a light bulb in a tiny cellar under the porch of my house. That bulb is on for a total of about ten minutes a year. For awhile, it looked as though I would be forced to replace the inexpensive incandescent bulb lighting it (at about 50 cents each) with an awful compact fluorescent light bulb (at a few dollars), back when there was talk of regulating all incandescent bulbs out of existance. Fortunately, cooler regulatory heads prevailed. At a total on-time of ten minutes annually, it’s doubtful that a CFL in my cellar would have paid back its additional cost in my lifetime, and I plan to live a long time. The cost equation for energy efficient lighting has improved since those days – you can find 60-W equivalent A19 LED bulbs for less than a dollar if you shop around. But I’m pretty sure most consumers would rather choose the kind of light bulb installed in their own home, rather than let some bureaucrat do it. The irony is LED bulb replacements for specialty incandescents are readily available now, no thanks to regulators. For example, I was able to find candle-shaped LED bulbs for less than three bucks each online.
2
DESIGN WORLD — EE NETWORK
2 • 2020
The 40-W incandescent versions seem to run around a dollar. If you’re curious about the payback time for these LED bulbs, we did a back-ofthe-envelope calculation using the 16.7 cents/kWh average power cost in California. It turns out that a 4.5-W, three-dollar LED candle bulb pulls even in cost with a 40-W, one-dollar incandescent version after about 336 hours of use at those rates. That works out to being switched on eight hours daily for about 42 days. Those economics will probably convince most consumers to make the switch to LEDs, with or without regulations in force. Thus the $2.4 billion that the Natural Resources Defense Council claims Californians will save on annual utility bills thanks to regulations would probably happen even if regulators sat on their hands. And that brings us back to the recently dropped lawsuit. I suspect NEMA and the ALA looked at prices of specialty LED bulbs, shrugged, and decided to forget the whole thing. Besides, California’s lighting efficiency regulations are completely unenforceable unless, of course, the State plans to somehow prevent online vendors from shipping incandescent bulbs to Californians. If that’s the plan, good luck.
LELAND TESCHLER | EXECUTIVE EDITOR eeworldonline.com
|
designworldonline.com
DESIGN WORLD FOLLOW THE WHOLE TEAM ON T WIT TER @DESIGNWORLD
EDITORIAL
PRINT PRODUCTION
MARKETING
EVENTS
VP, Editorial Director Paul J. Heney pheney@wtwhmedia.com @wtwh_paulheney
VP, Creative Services Mark Rook mrook@wtwhmedia.com @wtwh_graphics
VP, Digital Marketing Virginia Goulding vgoulding@wtwhmedia.com @wtwh_virginia
Events Manager Jen Osborne jkolasky@wtwhmedia.com @wtwh_Jen
Senior Contributing Editor Leslie Langnau llangnau@wtwhmedia.com @dw_3Dprinting
Art Director Matthew Claney mclaney@wtwhmedia.com @wtwh_designer
Senior Digital Marketing Strategist Mike Ulanski mulanski@wtwhmedia.com @wtwh_mike
Event Exhibitor & Speaker Manager Michelle Flando mflando@wtwhmedia.com
Executive Editor Leland Teschler lteschler@wtwhmedia.com @dw_LeeTeschler
Graphic Designer Allison Washko awashko@wtwhmedia.com @wtwh_allison
Digital Marketing Specialist Sean Kwiatkowski skwiatkowski@wtwhmedia.com
Executive Editor Lisa Eitel leitel@wtwhmedia.com @dw_LisaEitel
Graphic Designer Mariel Evans mevans@wtwhmedia.com @wtwh_mariel
Senior Editor Miles Budimir mbudimir@wtwhmedia.com @dw_Motion
Production Manager Barb Martin bmartin@wtwhmedia.com
Senior Editor Mary Gannon mgannon@wtwhmedia.com @dw_MaryGannon Associate Editor Mike Santora msantora@wtwhmedia.com @dw_MikeSantora
FINANCE Controller Brian Korsberg bkorsberg@wtwhmedia.com Accounts Receivable Specialist Jamila Milton jmilton@wtwhmedia.com
WTWH Media, LLC 1111 Superior Ave., Suite 2600 Cleveland, OH 44114 Ph: 888.543.2447 FAX: 888.543.2447
Director, Audience Development Bruce Sprague bsprague@wtwhmedia.com
VIDEOGRAPHY SERVICES Video Manager Bradley Voyten bvoyten@wtwhmedia.com @bv10wtwh
PRODUCTION SERVICES
Webinar Coordinator Halle Kirsh hkirsh@wtwhmedia.com
Customer Service Manager Stephanie Hulett shulett@wtwhmedia.com
Webinar Coordinator Kim Dorsey kdorsey@wtwhmedia.com
Customer Service Representative Tracy Powers tpowers@wtwhmedia.com
ONLINE DEVELOPMENT & PRODUCTION Web Development Manager B. David Miyares dmiyares@wtwhmedia.com @wtwh_WebDave Senior Digital Media Manager Patrick Curran pcurran@wtwhmedia.com @wtwhseopatrick
Videographer Derek Little dlittle@wtwhmedia.com @wtwh_derek
Front End Developer Melissa Annand mannand@wtwhmedia.com
Videographer Graham Smith gsmith@wtwhmedia.com
Software Engineer David Bozentka dbozentka@wtwhmedia.com
Event Marketing Specialist Olivia Zemanek ozemanek@wtwhmedia.com
Customer Service Representative JoAnn Martin jmartin@wtwhmedia.com Customer Service Representative Julie Ritchie jritchie@wtwhmedia.com
Digital Production Manager Reggie Hall rhall@wtwhmedia.com Digital Production Specialist Elise Ondak eondak@wtwhmedia.com Digital Production/ Marketing Designer Samantha King sking@wtwhmedia.com
DESIGN WORLD does not pass judgment on subjects of controversy nor enter into dispute with or between any individuals or organizations. DESIGN WORLD is also an independent forum for the expression of opinions relevant to industry issues. Letters to the editor and by-lined articles express the views of the author and not necessarily of the publisher or the publication. Every effort is made to provide accurate information; however, publisher assumes no responsibility for accuracy of submitted advertising and editorial information. Non-commissioned articles and news releases cannot be acknowledged. Unsolicited materials cannot be returned nor will this organization assume responsibility for their care. DESIGN WORLD does not endorse any products, programs or services of advertisers or editorial contributors. Copyright© 2020 by WTWH Media, LLC. No part of this publication may be reproduced in any form or by any means, electronic or mechanical, or by recording, or by any information storage or retrieval system, without written permission from the publisher. Subscription Rates: Free and controlled circulation to qualified subscribers. Non-qualified persons may subscribe at the following rates: U.S. and possessions: 1 year: $125; 2 years: $200; 3 years: $275; Canadian and foreign, 1 year: $195; only US funds are accepted. Single copies $15 each. Subscriptions are prepaid, and check or money orders only. Subscriber Services: To order a subscription or change your address, please email: designworld@omeda.com, or visit our web site at www.designworldonline.com POSTMASTER: Send address changes to: Design World, 1111 Superior Ave., Suite 2600, Cleveland, OH 44114
4
DESIGN WORLD — EE NETWORK
2 • 2020
eeworldonline.com | designworldonline.com
CONTENTS POWER ELEC TRONICS HANDBOOK | FEBRUARY 2020
6 20
38
32
02 06
WHY WE DON’T NEED ENERGY EFFICIENCY STANDARDS FOR LIGHTING
The need for high efficiency is pushing industrial power supplies toward synchronous topologies that minimize the need for heat sinking and cooling.
10
CAN A THREE-PHASE POWER SUPPLY OPERATE FROM WYE AND DELTA AC INPUTS?
14
THE BASICS OF GROUND-FAULT INTERRUPTION
20 24
28
THE MOVE TO SYNCHRONOUS POWER SUPPLIES IN SMART FACTORIES
A few guidelines help clarify what three-phase input voltage best suits a particular large power system.
Not all ground-fault interrupters are designed to protect living beings. Specialized industrial versions are optimized for maintaining the health of machinery and processes.
WHERE FIXED-RATIO CONVERTERS FIT IN HIGH-POWER DELIVERY SYSTEMS Fixed-ratio converters are often the best way of minimizing the electrical current sent over significant distances in the interest of better power efficiency.
COMPARING POWER SEMICONDUCTOR TECHNOLOGIES Cutting-edge power device technologies all have a niche. Here’s a rundown on the applications in which each one likely fits.
eeworldonline.com
|
designworldonline.com
28
HOT TOPIC: ELECTRIC VEHICLE COMPONENTS GET MORE SOPHISTICATED THERMAL MODELS
32
APPLYING LARGE BANKS OF SUPERCAPACITORS
38
THE TRY-BEFORE-YOU-BUY ROUTE TO ENERGY EFFICIENT POWER DESIGN
44
THE 40-YEAR BATTERY PACK
The high electrical currents of fast charging are forcing designers to invent standard techniques for predictably modeling heat and cooling.
It pays to know techniques for mitigating leakage current and over-voltages in uses where several supercapacitors work in parallel.
Evaluation platforms help pick SiC components in power conversion circuits.
Not all primary batteries offer the low annual selfdischarge rate needed for the lengthy lifespan of remote wireless devices consuming microamp-level currents.
2 • 2020
DESIGN WORLD — EE NETWORK
5
POWER ELECTRONICS HANDBOOK
The move to synchronous power supplies in smart factories ANTHONY “THONG” HUYNH
The need for high efficiency is pushing industrial power supplies toward synchronous topologies that minimize the need for heat sinking and cooling.
BUILDINGS, FACTORIES, AND INDUSTRIAL equipment are getting more space efficient, driving a move toward miniaturizing the underlying components that enable intelligence and automation. This trend challenges the power supplies for these designs. The power ICs must, of course, provide the necessary output voltages, and also must be small and capable of performing well in thermally difficult environments.
MAXIM INTEGRATED
There are several techniques and technologies that can ensure power supplies will meet the efficiency and size requirements of industrial equipment. They help realize the high efficiency, low power dissipation, and small form factor needed to address modern design targets. The task of reducing analog IC size hasn’t been easy—particularly because power management components consume a significant portion of PCB area. To illustrate this point, let’s examine the power supply requirements of some typical application areas, as well as what’s needed to address their demands.
Compact power circuits help trucking companies manage their fleets with vehicular asset-tracking devices. | JohnnyH5/iStock
6
DESIGN WORLD — EE NETWORK
2 • 2020
eeworldonline.com | designworldonline.com
SYNCHRONOUS POWER
Small, efficient power ICs support intelligence at the edge of automated factories such as this car manufacturing plant. | onurdongel/iStock
First consider vehicular asset-tracking devices. These aftermarket items let companies managing large vehicle fleets keep close tabs on their cars, trucks, and vans. Data collected can enhance fleet efficiency, ensure vehicles undergo regular maintenance, and locate vehicles in the event of theft or other issues. The devices themselves are usually powered by the vehicle battery (12 V for cars and 24 V for many trucks) and have a rechargeable backup battery. They’re typically installed under the dashboard, so heat dissipation is critical to keeping the device in its proper temperature range. The power circuit for asset-tracking devices can include step-down dc-dc converters and LDOs (low dropout regulators) to convert 12 and 24-V supplies to lower voltages for powering various digital logic and analog ICs. LDOs are, indeed, an option, as they’re typically easy to use and inexpensive. However, LDOs dissipate appreciable power when used directly from the main battery voltage, the primarily means by which these asset-tracking devices are powered. This application would be well served by highly integrated power ICs. Among today’s dc-dc regulators, there are devices that integrate the power MOSFETs, compensation circuit, and other external components, minimizing the discrete parts. In fact, new technologies push the envelope even further for efficiency and space savings. As intelligence continues to move to the edge of the digital factory, installations require a plethora of high-performance sensors, actuators, I/Os, and programmable logic controllers (PLCs). The insights they deliver allow factory equipment to become self-aware and deliver benefits such as greater productivity and up-time, predictive maintenance, and adaptive manufacturing. The power ICs for this equipment must be small, efficient, robust, and able to handle conditions such as drops, shock, and vibration.
eeworldonline.com | designworldonline.com
Industries that require noncontact level sensing--such as oil and gas, mining, and waste water processing--are driving demand for ultrasonic sensors. These types of sensors emit frequencies ranging from 30 kHz up to a few megahertz and interpret reflected signals as a sensing method. For level sensing, a single transducer is sufficient. After the transducer emits the ultrasonic signal, it measures reflection time from the surface of the liquid. Flow sensing typically relies on two transducers because this process uses the time-of-flight of the ultrasonic wave and its variation (depending on the flow rate of the medium). Ultrasonic sensors include an analog front-end that encompasses transmission, receiving, and control circuits, plus an MCU for the calculations. Increasingly, these sensors must be small enough to fit into compact quarters and in narrow openings. Their power supplies, therefore, must be small, rugged, and dissipate minimal heat. There are a few other power considerations to note. Fan-less ultrasonic sensors are typically powered by a supply of 10 to 30 V and consume a few hundred milliamps of load current. With system operating temperatures up to 70°C, the power components should be rated for at least 125°C. These systems also need protection from reverse polarity connections, over-voltages, and electromagnetic interference (EMI). Today’s smart buildings are defined by intelligence that controls the lighting, temperature, security systems, access mechanisms, and more. The sensors, controllers, and interfaces that make these capabilities possible are often tiny and, thus, require small, efficient
Industries such as waste water processing rely on ultrasonic sensors for level sensing. | BKhamitsevich/iStock
2 • 2020
DESIGN WORLD — EE NETWORK
7
POWER ELECTRONICS HANDBOOK Nonsynchronous buck converter
Efficiency comparison
VIN
EFFICIENCY (%)
95 90
SYNCHRONOUS BUCK
85
NONSYNCHRONOUS BUCK
IC CONTROL
80
70
0.5
1
L
MAX17506
VIN =24V VOUT = 5V
75
NONSYNCHRONOUS BUCK 1.5
2
2.5
OUTPUT CURRENT (mA)
3
3.5
D
power sources. Sensors generally receive power from 24-Vdc sources and must be able to operate in environments where there will be high-voltage transients. Non-critical industrial equipment typically demands a maximum operating range from 36 to 40 V. Critical equipment-which includes controllers, actuators, and safety modules--must support 60 V. On the output-voltage side, you’ll commonly find 3.3 and 5 V with currents from 10 mA in small sensors up to tens of amps in applications such as motion control. These parameters point to use of a stepdown voltage regulator that can withstand voltage transients (42 or 60 V, typically) for building and industrial control applications.
ENHANCING POWER SUPPLY EFFICIENCY Clearly power supplies for these applications share qualities that include small size, power efficiency, wide input-voltage range, and reliable operation in harsh and thermally challenging environments. Board space is DESIGN WORLD — EE NETWORK
often limited in these industrial designs, making it impractical to dissipate heat via techniques such as heatsinks. Forced-air fans also aren’t viable where sealed industrial enclosures prevent ingress of dust and pollutants. The only real option is a power supply with high efficiency. The input and output voltages involved in these industrial designs make it necessary to use a step-down, or buck, voltage regulator. The most common type of stepdown architecture is the nonsynchronous buck converter, with the low-side rectifier diode external to the IC. These devices are fairly easy to design for high voltages. In a design with a 24-V input and a 5-V output, the buck converter works with a duty cycle of approximately 20% and the external rectifier diode conducts the remaining 80% of the time (this accounts for most of the power dissipation). To greatly minimize the power dissipation, synchronous topologies replace the rectifier diode with a synchronous rectifier (such as a low-side MOSFET). Consider an example with a 4-A load 2 • 2020
VOUT COUT
4
The most common step-down architecture is the nonsynchronous buck converter partly because semiconductor manufacturers can easily design nonsynchronous buck regulators for high voltages. Here, the low-side rectifier diode is outside the IC. Synchronous architectures replace the diode with a low-side MOSFET acting as a synchronous rectifier. The voltage drop across the MOSFET’s T2 on-resistance, RDS(ON), is less than that of the diode the MOSFET replaces. Naturally, the goal is to fully integrate the entire synchronous rectification half-bridge (T1 and T2) into the IC, in a fully integrated synchronous converter. An efficiency comparison of the MAX17506, a synchronous step-down converter, versus a nonsynchronous design shows a higher efficiency for the synchronous solution across the higher load range. For both devices, the test conditions are 24 V input and 5 V, 4 A output. At full load (4 A), the efficiency of the synchronous solution is above 92% while that of the nonsynchronous device is only about 86%.
8
T
Synchronous buck converter VIN IC T1 CONTROL
L
T2
VOUT
COUT
Fully integrated synchronous buck converter VIN
IC T1 CONTROL
L
T2
VOUT
COUT
eeworldonline.com | designworldonline.com
SYNCHRONOUS POWER
Smart building applications such as automated lighting, security, and temperature control benefit from step-down voltage regulators that can withstand voltage transients. | zhudifeng/iStock.
and a Schottky rectifying diode that has a voltage drop of about 0.64 V. At 80% duty cycle, the conduction loss is roughly: (0.64 V) ×(4 A) × (0.80) = 2 W Replacing the diode with a low-side MOSFET serving as a synchronous rectifier replaces the 0.64-V drop with a drop across the MOSFET transistor’s on-resistance, RDS(ON). The MOSFET’s RDS(ON) is just 11mΩ. As such, the voltage drop is: (11 mΩ) × (4 A) = 44 mV And the power loss is: (0.044 V) × (4 A) × (0.80) = 141 mW In our example, the MOSFET power loss is about 14× smaller than the Schottky power loss at full load. Such are the power-efficiency benefits of synchronous rectification. A cautionary word about handling maximum input voltage: Though 24 V is the nominal rail for factory applications, it would be prudent to choose from the 28-V, 36-V, 42-V, or 60-V input power management modules now available. In fact, unless you know or will be able to model all possible surge scenarios that can stem from long cables and PCB traces, devices with a 42-V or 60-V maximum operating range should be the best bet—28 V is too close to 24 V to provide a reliable margin, and 36 V is risky when working with sensors and encoders on a 24-V rail. (This approach could expose equipment to excessive voltages, even with surge protection in place.)
eeworldonline.com | designworldonline.com
Integrated power modules can handle such demands while reducing overall circuit size by eliminating the need for many discrete components. For example, the Maxim Himalaya uSLIC dc-dc power modules integrate a synchronous wide-input Himalaya buck converter—including built-in FETs, compensation, and other functions—together with the output inductor. These modules reduce the power supply size by up to 2.25× compared to discrete approaches. The family of power modules supports input voltages from 2.9 to 60 V and, as such, can support low-voltage applications such as consumer devices as well as higher voltage industrial applications. All in all, integrated power modules help meet the input voltage, heat dissipation, and size requirements of the industrial applications that are bringing greater operational efficiency to our factories, vehicle fleets, buildings, and more.
REFERENCES Maxim Integrated Himalaya power modules, https://www.maximintegrated.com/en/ products/power/power-modules.html
2 • 2020
DESIGN WORLD — EE NETWORK
9
POWER ELECTRONICS HANDBOOK
Can a three-phase power supply operate from wye and delta ac inputs? TDK LAMBDA TECHNICAL MARKETING
A few guidelines help clarify what three-phase input voltage best suits a particular large power system. AC-DC POWER SUPPLIES rated higher than 2.5 kW frequently have a three-phase ac input. In the U.S. the voltage can be 208/220 Vac or 480 Vac. In Europe it is a “harmonized 400 Vac” which really is 415 Vac in the U.K. and 380 Vac for Europe. A higher input voltage allows more power to be drawn from the incoming ac at a lower
An HWS1800T-24 three-phase power supply.
current. These three-phase ac voltages can be one of two configurations – delta or wye (pronounced “why”). Typically high-voltage power is transmitted from the power generation plants to local substation transformers (where it is reduced in voltage) and then to facilities in a delta configuration. Note that a delta configuration only uses three wires and has no neutral or ground wire. This saves the cost of a fourth wire, which is unneeded during transmission. Consider the U.S. first. At a typical large power consumer such as a factory, a 480-Vac delta threewire feed enters the facility. (It comes to the facility from the local utility substation.) From the facility’s incoming distribution panel, 480 Vac delta is supplied
A TPS4000-24 three-phase power supply.
Genesys+ programmable power supplies GH1.5kW / G1.7kW
1ø 85 to 265Vac
G2.7kW / G3.4kW:
1ø 170 to 265Vac or 3ø 208Vac or 3ø 400Vac
G5kW / GSP10kW & 15kW: 3ø 208Vac, or 3ø 400Vac or 3ø 480Vac
10
DESIGN WORLD — EE NETWORK
2 • 2020
eeworldonline.com | designworldonline.com
THREE-PHASE POWER
3-phase delta power input from substation (480-Vac)
Power distribution in a major power user (US) Incoming distribution panel
Distribution panel
Phase A
Line 1
L-N Loads (120-Vac)
Neutral
L-L Loads (208-Vac)
Phase B Phase C
Line 2 Line 3 Step down Delta-Wye transformer Small 3-phase loads (208-Vac)
Large 3-phase loads (480-Vac)
A typical U.S. power distribution scheme for large power users such as factories locates a step-down delta-wye transformer on site. The transformer is generally large enough to sit on its own concrete pad, often outside the facility.
Power distribution in a major power user (Europe) Distribution panel Phase A High voltage (kV) power input from grid
Phase B Phase C
Line 1
L-N Loads (120-Vac)
Neutral
L-L Loads (208-Vac)
Line 2 Line 3 Step down Delta-Wye transformer 3-phase loads 400-Vac (380 / 415-Vac)
A typical European major power-using facility has its own delta-wye step-down transformer, like its U.S. counterpart, but the outputs are at different voltages.
directly to electrical equipment needing a large amount of power. Typical loads requiring this much power might include large ovens, test equipment for semiconductors, burn-in chambers, and machines fabricating metal (including laser cutting and additive manufacturing). It is important to note that connecting equipment to this incoming voltage feed, rather than just the reduced wiring gauge, can minimize the size of the delta-wye step-down transformer, cutting costs while saving energy and floor space. Power for the rest of the facility comes from a second distribution panel. This panel receives input power from a step-down delta-wye transformer that converts a 480-Vac delta configuration to a four-wire 208-Vac phase-to-phase wye configuration. The distribution panel, eeworldonline.com | designworldonline.com
in addition to being able to supply 208-Vac phase-to-phase, can also supply 120 Vac. The 120 Vac is available by connecting to either one of the output Lines (L1, L2 or L3) and neutral N. As a rough order of magnitude, 208 Vac three-phase would be used for mid-sized loads exceeding 5 kW but less than 25 kW. Single-phase 208 Vac is generally for smaller loads exceeding 1.5 kW. The 120-Vac wall outlet can support around 1 kW. The amount of power depends on the wiring size and fusing—consult your local qualified electrician! Some facilities may also contain a second delta-wye transformer. This transformer provides 277 Vac feeds to lighting and HVAC (Heating, Ventilation and Air Conditioning) equipment. And in Europe 2 • 2020
DESIGN WORLD — EE NETWORK
11
POWER ELECTRONICS HANDBOOK Phase relationship in wye output (US)
Phase relationship in wye output (Europe)
120Vac
Line L1
Line L1 208Vac
380 / 415-Vac
Neutral N - often grounded
Neutral N - often grounded
Line L3 120 Vac
Line L3
Line L2
120Vac
Lines coming off the wye step-down transformer connections measure 208 Vac phase-to-phase. Installations obtain 120 Vac by making connections between neutral and L1.
the arrangement and voltages are different than in the US. In European power distribution, the grid provides major power consumers with a high voltage (11 kVac in a delta configuration in the UK). A step-down transformer delivers three phase in a four-wire wye configuration to the facility’s distribution panel. Mainland Europe mainly uses 380/220 Vac while the U.K. uses 415/240 Vac.
Delta phase relationship
480Vac US Phase A
Phase B
Phase C
Delta wiring configurations measure 480 Vac phase-to-phase in the U.S.
12
DESIGN WORLD — EE NETWORK
220 / 240-Vac
In Europe the phase-to-phase wye output voltages are 380/415 Vac (Europe/U.K.) with 220/240 Vac available between L1 and neutral.
The distribution panel, in addition to supplying 380/415 Vac phase-to-phase, can also provide 220/240 Vac through connections to either one of the output Lines (L1, L2 or L3) and neutral N.
TYPICAL THREE-PHASE SUPPLIES For examples of three phase ac-dc power supplies that would work on the distribution systems outlined above, we shall review some examples from TDK-Lambda’s product offering. The HWS1800T-24 is a 1.8-kW-rated power supply accepting a 170-265 Vac three-phase input. This kind of input would be suitable for operation from a standard U.S.- type of 208-Vac three-phase wye input. It could also operate in Europe but would require a 400 -to-220 Vac three-phase wye-wye stepdown transformer. The TPS4000-24 is a 4-kW-rated power supply accepting a 350-528 Vac threephase input, either delta or wye. This supply would be suitable for operation in the U.S. and in Europe without changing connections to the power supply or additional transformers. The Genesys+ series of programmable power supplies encompasses a large number of models ranging from 1.5 kW to 15 kW. Depending on the power level, the 2 • 2020
Line L2
220 / 240-Vac
units have different input voltages covering most of the global input voltages. Regardless of the supply maker, there are a few points users of three-phase supplies should keep in mind. Ensure the manufacturer has internal fuses fitted, as some low-cost power supplies do not. A high-voltage fuse is required for each phase. They are bulky and are not inexpensive. With these points in mind, users of three-phase supplies might be tempted to take a second glance at those big grey mystery boxes – which you’ll now understand are step-down transformers — surrounded by chain-link fencing and high voltage warnings in the company parking lot!
REFERENCES TDK-Lambda Americas Blog, https://power-topics.blogspot.com/
eeworldonline.com | designworldonline.com
A57_DesWrld9x10_875.qxp_Layout 1 1/4/20 8:53 AM Page 1
The big name in miniature components. Transformers & Inductors, DC-DC Converters, AC-DC Power Supplies
The most demanding applications require the world’s most reliable components. For over 50 years PICO Electronics has been providing innovative COTS and custom solutions for Military, Commercial, Aerospace and Industrial applications. Our innovative miniature and sub-miniature components are unsurpassed in any industry. PICO Electronics’ products are proudly manufactured in the USA and are AS9100C Certified. To learn more about our products and how you can benefit from our expertise visit our website at picoelectronics.com or call us today at 800-431-1064.
think... low profile from
.18"
ht.
TRANSFORMERS & INDUCTORS
Surface Mount & Thru Hole
Think Pico Small Over 5000 std Ultra Miniature • • • • •
Military • COTS • Industrial
Ultra Miniature Designs MIL-PRF 27/MIL-PRF 21038 DSCC Approved Manufacturing Audio/Pulse/Power/EMI Multiplex Models Available For Critical Applications, Pico Continues to Be the Industry Standard
DC-DC CONVERTERS
2V to 10,000 VDC Outputs — 1-300 Watt Modules • MIL/COTS/Industrial Models • Regulated/Isolated/Adjustable Programmable Standard Models • New High Input Voltages to 900VDC • AS9100C Facility/US Manufactured • Military Upgrades and Custom Modules Available
PICO Electronics Inc.
143 Sparks Ave, Pelham, NY 10803 Call Toll Free: 800-431-1064 E Mail: Info@picoelectronics.com • FAX: 914-738-8225
VISIT OUR EXCITING NEW WEBSITE: www.picoelectronics.com
POWER ELECTRONICS HANDBOOK
The basics of ground-fault interruption WILL DELSMAN
NK TECHNOLOGIES
Not all ground-fault interrupters are designed to protect living beings. Specialized industrial versions are optimized for maintaining the health of machinery and processes.
Just a little current can kill
GROUND FAULTS ARISE when current flows from an energized conductor to ground inadvertently. The return path of the fault current is through living beings or equipment touching the
8000
grounding system. Ground fault detection is critical to protecting people and animals from shock or death.
• •
• •
•
•
•
14
At 1 mA you feel a slight tingle. At 5 mA you feel a slight shock, not painful but disturbing. The average individual can let go, but involuntary reactions can lead to injuries. At 6–25 mA there is painful shock, and muscular control is lost. The 9–30 mA level is called the freezing current or “letgo” range. At this level many humans cannot get their muscles to work, and they can’t open their hand to let go of a live conductor. At 50–150 mA there will be extreme pain, respiratory arrest, and severe muscular contractions. The individual cannot let go, and death is possible. At 1,000–4,300 mA there is ventricular fibrillation (the pumping action to the heart ceases). Muscular contraction and nerve damage occur. Death is most likely. At 10,000+ mA there will be cardiac arrest with severe burns and probable death.
DESIGN WORLD — EE NETWORK
2 • 2020
800 Current milliamperes
It doesn’t take much ground-fault current to cause harm. Extensive research in the 1960s determined the amount of current and voltage needed to cause ventricular fibrillation (where a heart stops beating) in humans. These studies found that as little as 70 mA through the heart was enough to cause fibrillation. The refinement of transistor technology provided a means of sensing currents as low as 0.003 A (3 mA) to energize a relay that would decouple the power supply. OSHA documents spell out the general relationship between the amount of current received and the reaction when current flows from the hand to the foot for just one second.
500 and higher Probably fatal
150
50 30 20 9 5 1
20-55 possibly fatal 30 standard Equipment protection 5-20 can’t let go Slight shock felt Barely feel tingle
eeworldonline.com | designworldonline.com
GROUND-FAULT INTERRUPTION Typical industrial ground-fault interrupter Ground fault sensor Sensor power
The current transformer in a ground-fault sensor detects any imbalance through the conductors of interest, measuring current leakage to ground. The sensor may then trigger a contactor to disconnect power or send an alarm to controls that initiate an appropriate action.
Electrical components be considered a UL943 GFCI. UL requirements for GFCI listing now include a self-test feature which must be included in the sensing and conditioning circuits.
Contractor EQUIPMENT ENCLOSURE Leakage to ground
Besides endangering lives, ground faults can also lead to costly fires and other equipment damage. Numerous safety regulations and electrical codes exist to prevent and protect against ground faults.
GROUND FAULT REGULATIONS The size of conductors, set points, and subsequent actions are all considerations for a ground fault sensor. What do the local codes require for protection and disconnect? What is the main goal in setting up a ground fault device? Is it focused on personnel protection or electrical device/process protection? In North America, ground fault circuit interrupters (GFCI) have been required by the National Electric Code (NEC) since the late 1960s. As the technology became more reliable, the NEC mandated GFCIs in many more applications to reduce the number of deaths from electrical shock. GFCI receptacles and circuit breakers were a huge step forward with a significant reduction in fatalities leading to a greater interest in ground fault protection. The NEC sets standards for where a GFCI is required and how quickly it should disconnect the circuit. The system design engineer follows the NEC Code requirements based on UL943. The UL specification designates what fault current level will cause the circuit to be de-energized and how quickly it must disconnect. NEC refers to personnel protection as ground fault circuit interrupters tested to UL943A requirements. A GFCI is designed to disconnect a circuit if current to earth exceeds 6 mA at 120 Vac. At a low level of fault current it may take a few seconds (UL943A states just under a minute) before the circuit is de-energized, while at a higher fault current (20 mA or higher) the circuit disconnects more quickly. Industrial ground fault sensors should be marked as recognized under UL1053, UL508 , or one of the other categories of UL943: subsection B, C or D. UL1053 is specific for ground fault sensing and relaying equipment with no stated current levels or time to operate. UL508 is an even broader category that covers a variety of automation components. When using a ground fault sensor to control a shunt-trip breaker, both components (sensing device and circuit breaker) must be tested together to verify the circuit is interrupted quickly enough to eeworldonline.com | designworldonline.com
PROTECTING PROCESSES While personnel safety is a major concern of ground fault protection, industrial settings dictate additional considerations. Manufacturing facilities normally employ a variety of safety protections for employees working with electrical machinery. Personal Safety Devices (PSD) used properly help minimize exposure to electrical dangers and allow fault current trip levels to be safely raised, minimizing nuisance trips and preventing undesired process interruptions. Ground fault detection can also be used to initiate controlled stops, alert other upstream processes, and even predict equipment failures. Underwriters Laboratories has established standards under UL943 for personnel protection (avoiding shock to humans) as well as for equipment protection at various fault levels and reaction time limits. The primary aim of equipment protection is to keep a fault from damaging machinery. Circuits supplying heating loads (heat strips, heat trace and snow melting equipment) are usually not disconnected until the fault current exceeds 30 mA or more. Electric vehicle charging stations are now required to have GFCI personnel protection according to 625.22.
Torrids in place
It’s possible to place a toroid over each conductor for sensing ground-fault current, then connect their secondaries in parallel. A sensing device then monitors the resulting circuit. However, this multiple-toroid method is inherently less precise than using a single toroid because of manufacturing and material tolerances.
2 • 2020
DESIGN WORLD — EE NETWORK
15
POWER ELECTRONICS HANDBOOK
Shunt trip with auto-rest ground fault Control power transformer
Out to monitored load (heating elements)
Shunt trip operating solenoid
DETECTION METHODS
Circuit breaker
A shunt-trip breaker lets a remote control open a circuit in the same manner as when there is an over-current condition. This action generally employs a magnetically operated solenoid to push or pull a latching mechanism to open the breaker contacts. The shunt-trip devices are also sometimes used to turn off loads in an emergency.
16
DESIGN WORLD — EE NETWORK
The 2017 NEC states the following: Article 426.28: Ground-fault protection of equipment shall be provided for fixed outdoor electric deicing and snow-melting equipment. Article 427.22: Ground-fault protection of equipment shall be provided for electric heat tracing and heating panels. This requirement shall not apply in industrial establishments where there is alarm indication of ground faults and the following conditions apply: 1) Conditions of maintenance and supervision ensure that only qualified persons service the installed systems, and 2) continued circuit operation is necessary for safe operation of equipment or processes. Article 555.3: The over-current protective devices that supply the marina, boat yards, and commercial and noncommercial docking facilities shall have ground-fault protection not exceeding 30 mA. The NEC calls for ground-fault protection for high-current supplies, too: Article 215.10 and 230.95 deal with current of 1,000 A and voltages of 480 or higher. Article 517.17 stipulates where fault detection is required in hospitals and other health care facilities. The importance of protecting an electrical system against faults-to-earth cannot be overstated. This type of fault sensing is not over-current detection, so fusing or circuit breakers will keep the conductors and insulation from being damaged.
2 • 2020
The primary method of fault detection utilizes a single ring of magnetically permeable material wound with many turns of small gauge wire (forming a current transformer or toroid) surrounding all the current-carrying conductors. If there is more current supplied to the load than is returned to the source, this sensing toroid produces a low voltage in the windings. This voltage is amplified and used to trigger an action such as energizing a solenoid to open a set of contacts. It is also possible to place a toroid over each conductor connecting the secondaries together with a sensing device installed to monitor the resulting circuit. Each toroid or current transformer must be rated to handle the maximum current in each conductor. The accuracy of this multiple toroid method is inherently less precise than using a single toroid due to manufacturing and material tolerances. A similar approach can be used to monitor the entire load of a machine or distribution panel supplied by a wye-connected (star) transformer and bonded to earth at the machine location. Passing only the bonding conductor through a ground fault sensor will perform the same function as using a large toroid over all the conductors. In most industrial applications, the ground-fault sensor output performs one of two operations: A contact closes a circuit to energize the operating solenoid of a shunt trip circuit breaker or, a contact opens a circuit powering a contactor or motor starter operating coil. How the sensor output interacts with the rest of the control system is completely at the discretion of the system designer. Circuit breakers come in a variety of styles. Some can accept feature-enhancing accessories (such as for under voltage trip, alarm contacts, and interchangeable trip plugs). Most common is the shunt-trip breaker that allows the circuit to be opened from a remote eeworldonline.com | designworldonline.com
GROUND-FAULT INTERRUPTION Commercial kitchen application
Grill
Neutral bar
Shunt trip operating solenoid Circuit breaker
location (acting as if there was an overcurrent condition). This action is commonly accomplished by a magnetically operated solenoid that pushes or pulls a latching mechanism to open the breaker contacts. In some applications the shunt-trip device turns off a load in an emergency. For example, most auto fuel dispensing stations have an emergency switch that disconnects all fuel pumps if there is a problem. This switch closes the circuit to operate a shunt-trip breaker, removing power from the pumps. The circuit breaker must be reset manually once the fault condition has been addressed. When a shunt-trip is used with an autoreset ground fault sensor, the sensor contact closes the circuit to the shunt solenoid when it detects a fault over the sensor trip point. As in other cases, the breaker must be reset manually after the cause of the fault is determined and mitigated. Because power eeworldonline.com | designworldonline.com
In one commercial kitchen application, designers used an NK Technologies Tri-Set ground-fault sensor with a range jumper installed at the highest setpoint (30 mA), allowing the equipment to operate during its initial burn-in. With the burn-in complete, the sensors were readjusted to a 5-mA setpoint. An example of a sensor that works well for such uses is the AGLD.
to the monitored load is turned off, and the only way to restore power is to reset the circuit breaker. Using a shunt-trip accessory effectively transforms an automatically resetting sensor into a latching device. Even if the cause of the fault is removed from the load and the sensor remains powered from an isolated source (recommended for all installations), the load cannot be energized until the breaker is reset. A latching output sensor, like the autoreset models, is typically equipped with an integral test button. Two additional terminals allow attachment of an external contact, usually a button mounted to the enclosure door, enabling the sensor to be reset after a fault is detected without opening the panel. Another common method used with a ground fault sensor is to have the contact open the circuit providing power to a contactor coil, de-energizing the load – typically multiple heating elements or a motor-driven pump or fan. Opening a contact in a control system sounds easy, but in most ground-fault sensing applications, the contact must be closed before the monitored load is energized. Manufacturers offer both normally energized and normally de-energized versions of auto-reset ground-fault sensors. The more common of the two is normally deenergized in which the output, whether solidstate or an electromechanical relay, does not 2 • 2020
change state unless there is a fault-to-ground exceeding the trip point. The normally energized version is sometimes referred to as fail-safe. Here, the output changes state when the sensor first powers up. The output returns to normal or “shelf-state” condition when one of two things happen: the sensed fault current exceeds the trip point, or the power to the sensor is removed. In the case where a normally open, normally energized solid-state output model opens the circuit powering a contactor coil, the output contact would be open at shelf state and closed when the monitored circuit is not passing current to ground and the sensor is energized. The sensor output will open, turning off the monitored load, if the sensor power is interrupted or if the monitored load passes current to ground exceeding the trip point. It is important to understand that the monitored circuit might not energize if the sensor did not see power first, as energizing the sensor closes the output contact. More commonly the sensor selection would be normally closed, normally de-energized (solid-state) with the contact opening only when current exceeding the trip point passes to ground. With electromechanical relay outputs the operation is the same. In normally energized versions, the output relay is energized with sensor power applied so the contacts change state when the sensor has power. The relay will DESIGN WORLD — EE NETWORK
17
POWER ELECTRONICS HANDBOOK
then return to shelf state when there is a loss of power to the sensor or the fault current exceeds the trip point. When an auto-reset sensor output controls a contactor, it’s best to use a threewire connection method (like a standard momentary motor-start/stop-button setup) so the contactor must be re-engaged after the sensor trips. Alternatively, a latching-output version of the sensor is an option. In some code jurisdictions a contactor might not be considered a circuit disconnect. The local inspector, specifier, or AHJ (authority having jurisdiction) has the final say. The best place to monitor a circuit for ground faults is close to the load rather than a distance upstream. Many system designers tend to specify sensors that will monitor several loads simultaneously by installing the sensor before a final distribution point. The problem is that any minor leakage in each load accumulates, resulting in a higher leakage current level overall. As an example, visualize a machine that produces silicon wafers for electronics. Several heating elements warm chemical wash processes, several motors perform product positioning, and there are transformers adjusting voltage levels for various process controls. Sensors can be set to trip at relatively low levels if the motors, transformer loads, and heating elements are monitored for faults individually. But if a single sensor protects all loads, the trip point likely must be set much higher, reducing the
protection level of each piece of equipment. Heating elements seldom leak low-level current the way motors and transformers do. In most cases when heaters fail, there is a direct short-to-ground or the circuit is completely open. Heat trace cable runs do tend to leak small amounts of current to earth or there may be capacitance losses in long runs. With loads such as motors and transformers, small imperfections in the varnish insulation of the windings can let low levels of current pass to earth. While humans can seldom feel 3 or 4 mA currents, this low current leakage can rise over time until it becomes a concern both to personnel and the equipment itself. The ability to precisely monitor groundfault leakage lets the operator decide how to handle any ground-fault conditions. In applications where deterioration over time is expected, the monitoring of ground leakage levels can determine specific maintenance or replacement needs and prevent costly unexpected equipment failure and shutdowns. There are numerous applications where leakage-to-ground can exceed 30 mA yet not cause harm. And in some circumstances, disconnecting power prematurely may cause significant machine or process damage. Environmental conditions, such as excessively humid or wet conditions caused by washdown, or failing enclosure seals may elevate leakage. In applications where actions are required at predetermined or specified leakage levels, the factory calibrated setpoint will simplify setup. Once the setpoint is established, the designer chooses how to specifically handle the fault. Local codes may determine the sequence of events once a fault occurs. Environments characterized by widely ranging temperature and moisture conditions wreak havoc on electrical systems. Changes in heat and humidity eventually break down protective insulation
to cause ground leakage. Wet environments pose additional concerns as potential shock hazards multiply. Unlike moisture sensors that must be wired back to the motor control center, the ground fault sensor installs directly in the control panel, minimizing wiring. Industrial electrical heaters are prone to ground leakage from the breakdown or contamination of insulation. The on/off output of the ground-fault sensor can be used to trigger a circuit interruption device (such as a shunt-trip breaker) or a monitoring device (like a PLC) to determine the required action.
SPECIAL SITUATIONS It can be useful to review how groundfault equipment has served in particular applications that each have unique needs. For example, recent updates to NEC Code 555 require that marina owners consider ground-fault protection at both the individual slips and at the power distribution center feeding the separate branches to each slip’s power pedestal. Typical power feeds now require sensors that can handle conductors carrying more than 300 A, necessitating use of additional components. To address this problem, NK design engineers developed a sensor with an aperture measuring four inches in diameter allowing conductors (carrying 800 A or more) to easily pass through the sensor. These large-aperture AG-LC sensors can monitor the main circuit feeding the pedestals and energize a shunt trip breaker protecting the entire docking facility. Smaller aperture sensors can monitor individual power pedestals at each slip, with the sensor output energizing a shunt-trip breaker at the pedestal. In the case of kitchens, NFPA NEC 2017 Sec. 210.8 requires GFCI for personnel protection in commercial kitchen equipment with “single-phase receptacles rated 150 V to ground or less, 50 A or less, and threephase receptacles rated 150 V to ground
An example of a ground-fault sensor designed for use for industrial settings is the AGT model which has a 0.74-in (19 mm) aperture. It responds withing 600 msec to a 90% step change in current and puts out an analog signal proportional to the sensed current.
18
DESIGN WORLD — EE NETWORK
eeworldonline.com | designworldonline.com
GROUND-FAULT INTERRUPTION
Marina application
Ground fault detection at main feed location — transformer primary Main transfromer to boat slips Ground fault detection at main feed location — transformer secondary
To power pedestals 120v 240v
Ground fault detection at each power pedestal
Power pedestals 120v 240v
Ground fault detection at each power pedestal
To handle new code requirements for marinas, NK Technologies design engineers developed a sensor with an aperture measuring four inches in diameter. This allowed conductors carrying 800 A or more to easily pass through the sensor. The large-aperture AG-LC sensors can monitor the main feeding circuit to power pedestals at each slip and energize a shunt-trip breaker protecting the entire docking facility.
eeworldonline.com | designworldonline.com
or less, 100 A or less.” Prior to this change, only 15- and 20-A single-phase circuits of 125 V or less needed this level of protection. Circuit breakers and receptacles meeting this requirement are readily available and quite common. Requirements above 20 A or needing three-phase protection are a more difficult issue. Additionally, commercial kitchen steamers and grills sometimes retain humidity while stored prior to installation. So units must be “burned in” or energized for at least two hours before normal use. The additional moisture present during this process increases the ground-fault leakage to a point above the 5-mA trip level. To avoid nuisance tripping during the burn-in cycle, a ground-fault sensor must allow a temporary rise in the setpoint. To handle such situations, some sensors offer adjustable capabilities as a standard feature. A factory placed range jumper is installed at the highest setpoint (30 mA) allowing the equipment to operate during the initial burn-in. With the burn-in complete, the sensors can be readjusted to the 5-mA setpoint. The fabrication of silicon wafers into semiconductor chips involves hazardous chemicals and extreme heat. The SEMI standard S22-071b provides guidelines regarding the safety of semiconductor processing equipment, including Emergency Mains Off (EMO) circuitry design. This standard requires a means for the operator to easily disconnect mains power should any problem arise during processing. Because there are electrical heating elements throughout the fabrication equipment, ground fault protection is paramount. The elements are monitored in each process segment, and fault detectors are set at fairly low trip points. If there is a faultto-earth through the heating element, sensors selectively shut down only the affected part of the process. If several heating processes short simultaneously, a sensor with a bit of delay and higher trip point shuts off the main power feed. Here, sensors with adjustable setpoints and delays help manage the controlled shut down of the system in case of critical failures. It is relatively easy and quite common to detect low-level current in ac circuits. In North America, codes require all electrical outlets 2 • 2020
mounted in wet conditions be protected with ground-fault circuit interrupters. If ac current of 4 to 6 mA passes to ground, a circuit breaker or the contacts in the power receptacle open before there’s an electrocution. Most electrical heating elements must also be protected to keep equipment from damage in the event of a fault. In contrast, trying to detect the same fault condition in a dc circuit with a floating ground is not as simple. With the proliferation of photovoltaic panels and other alternative power sources, the need for ground-fault detection in dc-powered systems is critical. Solar panels or battery-operated systems use positive and negative conductors that are insulated. When connections get wet, this insulation becomes compromised and current can pass to earth. Water is the most common cause of dc fault current, while deteriorating insulation and contaminants on battery housings are additional factors. Because dc current leakage to earth presents a dangerous situation, early fault detection is essential. Fault detection that doesn’t add impedance to the monitored circuit is the safest approach.
REFERENCES NK Technologies, www.nktechnologies.com
DESIGN WORLD — EE NETWORK
19
POWER ELECTRONICS HANDBOOK
Tethered drones are an example of a situation where power cables must be as light as possible. High-voltage converters can help reduce the size and weight of the tethering.
Where fixed-ratio converters fit in high-power delivery systems Fixed-ratio converters are often the best way of minimizing the electrical current sent over significant distances in the interest of better power efficiency.
THE VAST MAJORITY OF electromechanical or
PHIL DAVIES
VICOR CORP.
and regulation requirements.
Many power system designers consider regulated dc-dc converters as essential to their overall systems design. However, PDN regulation is not always necessary for providing the right level of voltage to the PoL regulators nor imperative to an intermediate distribution-bus voltage. With this in mind, power system engineers should consider implementing fixed-ratio dc-dc converters, which can offer significant advantages to the overall performance of the PDN. PDN performance is commonly measured in terms of power loss, transient response, physical size, weight and cost. One major design challenge impacting PDN performance is the number of times the network needs voltage conversion and tight line/load regulation. Engineers spend a great deal of time optimizing bulk power voltage conversion, dynamic regulation and distribution qualities to deliver high performance and reliability. If system load power is in the multi-kilowatt range, designing the bulk PDN to handle a high voltage reduces the current the system must distribute (P= V×I). Consequently the PDN size, weight, and cost (cables, bus bars, motherboard
20
2 • 2020
semiconductor loads require stable dc-dc voltage conversion and tight regulation to operate reliably. The dc-dc converters that perform this function are commonly called point-of-load (PoL) regulators and are designed with a maximum and minimum input voltage specification defining their stable operating range. The power delivery network (PDN) to these regulators can vary in complexity based on the number and type of loads, overall system architecture, load power levels, voltage levels (conversion stages), as well as isolation
DESIGN WORLD — EE NETWORK
eeworldonline.com | designworldonline.com
FIXED-RATIO CONVERTERS
Paralleled bus converters copper power planes) can be reduced (PLOSS = I2R). Thus designers strive to keep as much circuitry as possible operating at highvoltage/low-current, only converting to low-voltage/high-current close to the load. However, bringing a high-voltage and high-power PDN close to the load requires a dc-dc converter with high efficiency and high power density. If the circuit demands a large step-down in voltage, as from 800 or 400 V to 48 V, the converters able to do the job and having the highest efficiency would be fixed-ratio converters. These converters provide no regulation and dissipate little power. Their high efficiency figure-of-merit enables higher power density and easier thermal management.
PC
+IN
eeworldonline.com | designworldonline.com
+IN
+OUT
–IN
–OUT
+OUT
PC L_2
BCM_2
+IN
+OUT
–IN
–OUT
PC
WHAT IS A FIXED-RATIO CONVERTER? A fixed-ratio converter operates much like a transformer but instead of ac-ac conversion, performs dc-dc conversion with the output voltage being a fixed fraction of the dc input voltage. As with a transformer, the converter provides no output voltage regulation, and the input-to-output voltage transformation is defined by the “turns ratio” of the device. This turns ratio, referred to as the K factor, is expressed as a fraction relative to its voltage step-down capability. K factors can range from a K = 1 to as low as K = 1/72 and are selected based on PDN architecture and the PoL regulator design specifications. Typical PDN voltages are categorized as low voltage (LV), high voltage (HV) and to ultra-high voltage (UHV). Fixed-ratio converters can be isolated or non-isolated and also capable of bidirectional power flow with reverse voltage conversion. For example, a K = 1/16 fixed-ratio converter with bidirectional capability can be operated as a boost converter with a K of 16/1. Additional design flexibilities include easy paralleling to meet higher power demands and the option of connecting converter outputs in series to boost output voltages by, in effect, changing the K factor. Power delivery networks are undergoing significant changes due to the soaring power demands within many end markets and applications. EVs (electric vehicles), mild hybrid and plug-in hybrid vehicles are using higher PDN voltages such as 48 V. The 48-V level meets the SELV (Safety Electrical Low Voltage) standard required by many systems, and the simple power equations of P = V×I and PLOSS = I2R explain why higher-voltage PDNs are more efficient. For a given power level, the current is four times lower at 48 V than in a 12-V system and has losses 16 times lower. At one-quarter of the current, the cables and connectors can be smaller, weigh less and be inexpensive. The 48-V battery used in hybrid vehicles has four times the power of a 12-V source, and the added power can be used in powertrain applications to reduce CO2 emissions, improve gas mileage, and handle new safety and entertainment features. The addition of AI (artificial intelligence) in data centers has driven rack power dissipation above 20 kW which has made the use of a 12-V PDN bulky and less efficient. Use of a 48-V PDN brings the same benefits here as with hybrid vehicles. In both automotive and data center applications, the preference is to keep 12-V legacy
BCM_1
L_1
L_N
BCM_N
+IN
+OUT
–IN
–OUT
–IN
–OUT PC
BCM converters are easy to parallel to meet higher power demands.
Bus converters in series PC
BCM_1 +IN
+IN
+OUT
–IN
–OUT
+OUT
PC
BCM_2
+IN
+OUT
–IN
–OUT
PC
BCM_N +IN
+OUT
–IN
–OUT
–IN
–OUT BCMs with outputs connected in series for higher output voltage enable more design flexibility.
2 • 2020
DESIGN WORLD — EE NETWORK
21
POWER ELECTRONICS HANDBOOK Going from 48 to 12V The Vicor NBM2317 enables efficient conversion from 48-to-12 V and vice versa as it is a bidirectional converter. The bi-directionality enables integrating a legacy board into a 48-V infrastructure or the latest GPU into a legacy 12-V rack.
loads and PoL commodity buck regulators to minimize the amount of change. Because 48 V is SELV-compliant, a non-isolated fixed-ratio converter is a good choice for the 48-to-12-V dc-dc conversion stage, as today’s PoL 12-V regulators can handle the variation in input voltage. A non-isolated, unregulated fixed-ratio converter is the most efficient high-power bus converter. It lowers power dissipation, boosts power density, and cuts costs. Its high density allows new decentralized distributed power architectures in hybrid automobiles where non-isolated fixed-ratio converters can sit near the loads, making possible smaller and more efficient 48-V PDN wiring throughout the vehicle. In server blades, a small non-isolated 48-to-12-V fixed-ratio converter can sit on the motherboard close to the buck regulators. Many new AI accelerator cards, such as the SXM from Nvidia and the OAM cards from Open Compute Project (OCP) members, are designed with a 48-V input because the AI processors consume 500 – 750 W. Cloud computing and server companies still using 12-V PDN backplanes in their racks require 12-to-48-V conversion to employ these high-performance cards. Equipping these accelerator cards with a bidirectional K = 1/4 non-isolated fixed-ratio converter acting as a 12-to-48-V boost converter (K = 4/1)—or within a distributed higher power 12-to-48-V module—enables older rack systems to incorporate AI capability.
Typical EV 48-V system A decentralized 48-V architecture places multiple smaller, lower-power converters closer to the 12-V loads.
22
DESIGN WORLD — EE NETWORK
2 • 2020
HIGH-VOLTAGE APPLICATIONS THAT REQUIRE ISOLATION The world is now equipped with 4G radio and antenna towers which must be upgraded with new 5G systems consuming 5x the power of 4G equipment. The 4G PDN is 48 V and is delivered via cable from a ground-based power system. The significant rise in power consumed by 5G equipment would force use of a large-diameter and heavy power cable were the PDN to stay at 48 V. So telecommunications companies are currently looking at using a 380-Vdc PDN to significantly reduce the cable size. Use of a bidirectional K 1/8 fixed-ratio converter in boost mode lets the ground-based 48-V power system deliver 380 V (K: 8/1) to the top of the tower. A 380-to-48-V regulated converter at the top of the tower will allow both 5G and 4G systems to receive a regulated 48-V supply and realize less expensive power delivery via a small 380-V power cable. Tethered drones are another high-voltage application that requires isolation. Power cables for tethered drones can be over 400 m long, and the drone must lift this cable weight when flying. Use of a high voltage such as 800 V helps reduce the size and weight of the tethering power cables. A compact onboard fixed-ratio converter, typically K = 1/16, can step power down to 48 V for the onboard electronics and video payloads. In EVs, high power demands make 400 V a common choice for battery voltage. The 400 V is then converted to 48 V for distribution to the various loads around the powertrain and chassis. For fast-charging, the 400-V battery is charged from a charging station having a regulated 800-V dc output via an 800-to-400-V converter. In both the 400/48-V and 800/400-V applications, a parallel array of isolated K: 1/8 (400/48) and K: 1/2 (800/400) fixed-ratio converters with high power density and efficiency
eeworldonline.com | designworldonline.com
PowerShuntAd_DesignWorld2_20.qxp_Layout 1 1/10/20 4:42 PM Page 1
FIXED-RATIO CONVERTERS
Power Shunt
Current Sense Resistors
Fixed-ratio converter K=1/16
High-voltage side
Low-voltage side
K=16/1
A bidirectional fixed-ratio converter operating as a step-down converter with K = 1/16 can also serve as a boost converter with a K of 16/1.
above 98% can work effectively. Regulation comes either before or after the fixed-ratio converter stage. The power density and efficiency gains of not having regulation also simplifies thermal management. Exascale High Performance Computing (HPC) systems use 380 Vdc as the main PDN because rack power levels typically exceed 100 kW. In these applications, isolated fixed-ratio converters of K: 1/8 and K: 1/16 are integrated onto the server blades or on mezzanine cards distributed through the rack to deliver either 48 V or 12 V to the motherboards. Regulation then comes via a 12-V multiphase buck converter array or advanced higher-efficiency 48-V-to-PoL architectures. The density and efficiency of the fixed-ratio converter again plays a critical role in enabling this type of PDN architecture to deliver high performance. Advanced systems in enterprise and high-performance computing, communications and network infrastructure, autonomous vehicles and numerous transportation applications are just a few of the high-growth industries clamoring for more power. These applications have a common thread: Each has extreme power requirements and benefits from a small, power-dense dc-dc converter that can save space and weight. Power system engineers should consider fixed-ratio converters as an important and flexible way of enabling high-performing PDNs that provide a competitive advantage in overall system performance.
Concept to Reality
KOA Speer’s continually expanding line of Power Shunt Current Sense Resistors are the ideal solution to help you optimize your power system design. Our new Power Shunt Current Detectors are designed for DC to DC converters, inverters, batteries, motor controls, automotive modules, power supplies or any other power management application. KOA Speer Power Shunts Deliver: • High Power: up to 10W • Ultra Low Resistance: 0.5mΩ ~ 1mΩ • 4 Terminal, 2726 size - PSG4 2 Terminal, 3920 size - PSJ2 • Wide Temp Range: -65°C ~ +175°C
Check out the new products from the industry’s most award winning passive component supplier… KOA Speer.
Visit KOASpeer.com
REFERENCES Vicor Corp., www.vicorpower.com Vicor NBM2317, http://www.vicorpower.com/industries-and-innovations/nbm
DESIGN WORLD — EE NETWORK
Help you Make the Leap from
23
POWER ELECTRONICS HANDBOOK How wide bandgap technologies stack up
A comparison of Si, SiC and GaN material qualities.
Comparing power semiconductor technologies DR. GERALD DEBOY, DR. PETER FRIEDRICHS Cutting-edge power device technologies all have a niche. Here’s a rundown on the applications in which each one likely fits.
INFINEON TECHNOLOGIES
THE PROCESS OF converting electrical power to different ac and dc levels and to mechanical motion has changed dramatically over the years. The development of “switched-mode” techniques has been among the most disruptive changes. Switched-mode techniques exponentially boost power conversion efficiency and make practical developments such as isolated dc-dc conversion without motorgenerator sets. Semiconductor switches are, to a large degree, what makes switchmode conversion possible. Today we have a wide range of power switches from which to choose. They include IGBTs (insulated-gate bipolar transistors) and silicon MOSFETs, as well as wide-bandgap devices based on both silicon carbide (SiC) and gallium nitride (GaN). The application determines which technology may be more reliable, efficient and economical. Design engineers choosing a power switch must evaluate each device’s operating range, properties, challenges and benefits, as well as specific application needs.
24
DESIGN WORLD — EE NETWORK
2 • 2020
eeworldonline.com | designworldonline.com
POWER SEMICONDUCTOR TECHNOLOGIES Where power technologies excel The power handling and switching frequency ranges now predicted for power switches technologies.
SEARCHING FOR THE IDEAL SWITCH Power converters are becoming a focal point for saving energy, lowering costs and minimizing environmental impact. Highefficiency converters also bring smaller, lighter and cheaper products without a power consumption penalty and deliver more power in the same volume for the same temperature rise. Power density is also an important metric in applications such as cellular infrastructure, EV/HEVs and edge computing. When evaluating power switches, many believe wide bandgap semiconductors based on SiC and GaN are ideal. GaN seems to be the clear winner in applications requiring a 650-V blocking voltage while SiC is superior above 1,000 V. However, practical components have several qualities that may affect their performance in specific uses. In many applications the latest-generation Si MOSFETs, with lower blocking voltages and high-voltage IGBTs, may cost less and perform better. Power converters use four main switch types, Si IGBTs, Si MOSFETs, SiC MOSFETs or GaN HEMT (high-electron-mobility transistor) cells. In addition, there are hybrids of the technologies. Examples include IGBTs with integrated parallel SiC Schottky barrier diodes or IGBT and SiC switches combined
eeworldonline.com | designworldonline.com
in multilevel topologies, like NPC or ANPC (neutral-point-clamped and active neutralpoint-clamped). Here each component excels at different points in the topologies. GaN is best for new designs. Simply replacing IGBTs or Si MOSFETs with GaN devices and adjusting the gate drive probably won’t bring benefits and may just increase EMI. But existing topologies may see benefits from replacing Si IGBTs with SiC. There can be significantly less loss even with no change in either switching frequency or turn-on/turnoff times. In some applications, the device’s bidirectional capability, or how it operates in its so-called “third quadrant,” is critical. This mode of operation arises when both the drain voltage (collector for IGBTs) and current are negative in n-channel devices, a condition happening under “commutation” in halfbridge networks. This operation mode appears in many common circuits such as inverters, motor drives, totem-pole PFC stages, and the now widely used LLC, phase-shifted bridge and active-clamp flyback converter topologies. Si and SiC MOSFETs have integral body diodes that conduct under these negative current/voltage conditions. These diodes also exhibit a varying voltage drop and reverse
2 • 2020
recovery currents of differing amplitudes and durations. However, IGBTs have no intrinsic diode. As the device switches off, a current “tail” causes high power dissipation. Si MOSFET body diodes have a high reverse recovery charge and are relatively slow, causing losses especially at high frequencies. SiC MOSFETs have negligible reverse recovery charge, but the diode has a high forward-voltage drop when it conducts, again dissipating energy. Thus, designers increasingly use synchronous rectification (replacing diodes with actively controlled switches) which, in turn, lowers static losses below that when diodes are used in combination with IGBTs. On the other hand, GaN conducts in the third quadrant with a drop defined by low channel resistance and with no reverse recovery at all. Both effects help minimize power losses. GaN and SiC are, therefore, ideal solutions for optimizing efficiency with half-bridge-based topologies, particularly under “hard” switching conditions. With “soft” or resonant switching, Superjunction Si MOSFETs are still the optimal choice up to about 250 kHz. Current and voltage ratings also limit GaN applications, making IGBTs and SiC the dominant technologies for higher currents and voltages.
THE IMPORTANCE OF POWER DENSITY Although efficiency is key, high power density is sometimes the actual design aim. The maximization of power density involves a choice between high-frequency operation with GaN, potentially reducing the size of passive components, or operating Si or SiC MOSFETs at lower frequencies with slightly larger passive components. SiC and GaN circuitry, with typically better efficiency, allows use of smaller heatsinks. But the thermal design of associated components must be done with care. The selection of topology and control strategies for hard or soft switching (where the voltage or current through the device is zero when it switches) also play important roles.
DESIGN WORLD — EE NETWORK
25
POWER ELECTRONICS HANDBOOK Energy in power devices
A power converter’s total cost of ownership (TCO) will include not only acquisition cost and energy wasted through inefficiency, but also the amortized design, development, and qualification expenses along with the cost of scraping obsolete equipment. Higher power-converter efficiency speeds payback especially in energy-hungry applications such as hyper-scale data centers. For example, consider the energy efficiency efforts of the Open Compute Project, where OCP members share designs for data center products. The trend is to integrate power converters into the rack, with projected lifetimes of six to eight years, instead of the typical two-to-three-year life of power supplies mounted directly on motherboards. At an efficiency level close to 98%, GaN designs pay for themselves after around three years depending on electricity costs. Their reduced converter size and higher efficiency can also allow more server blades to fit in each rack, boosting data throughput. The same benefits are available in other applications. In motor drives, for example, more efficient GaN technology could substantially reduce inverter temperature rise and allow twice as many drives to sit in a cabinet, greatly reducing the need for expensive factory floorspace.
FIGURES OF MERIT The figure of merit (FOM) for a power-handling devise is the product of ON-resistance and die area – RDS(ON)×A. FOM is a good indication of how much power the device handles for a given voltage rating. A low value implies a smaller die with correspondingly smaller device capacitances and higher switching speeds. It also implies more die per wafer for a given power rating, potentially lowering device costs.
26
DESIGN WORLD — EE NETWORK
2 • 2020
The relative performance of IGBTs, Si Superjunction MOSFETs, SiC MOSFETs and GaN HEMTs (with and without losses in the RC circuit) from Infineon in terms of gate charge and driving power (inset).
A similar FOM given by RDS(ON)×EOSS combines conduction losses with switching loss generated by the dissipation of stored energy, EOSS, in device output capacitance each switching cycle. The lower the FOM, the lower the losses. EOSS itself, the energy stored in device output capacitance, is a similar indicator of turnon losses. Total gate charge QG(TOT) affects gate drive power PG required, according to: PG = QG(TOT) × F × VGSW where F = frequency, Hz; and VGSW = gate voltage swing, V. This relationship is important because gate-drive power can affect system efficiency. GaN has a low threshold voltage of around 1.5 V and extremely low QG(TOT). So gate-drive power is just milliwatts even at high frequencies. On the other hand, IGBTs can require several watts of gate drive, even at low frequencies, because of their typical gate-voltage swings between +16/-9 V and gate charge measured in microcoulombs. SiC lies between the two. The charging and discharging of the capacitance associated with QG(TOT) through device and external gate resistance also affects switching speed and EMI generated. The markets are growing for Si, SiC and GaN alike. Analysts at Research and Markets, for example, expect the market for IGBTs to grow at a CAGR of more than eight percent and amount to nearly US$10 billion by 2023. Si MOSFETs are still preferred for certain applications. Potential applications for wide bandgap semiconductors continue to expand, so active development of new products with improved specifications continues across the board. To get a feel for the direction of the trend, consider work underway at Infineon. Infineon aims to move towards ever-better RDS(ON)×EOSS eeworldonline.com | designworldonline.com
POWER SEMICONDUCTOR TECHNOLOGIES Performance roadmap How Infineon Si-Superjunction MOSFETS have improved over time.
and RDS(ON)×A. Infineon pioneered Si-Superjunction technology and expects to halve the RDS(ON)×A FOM to about 5 mΩcm² in the future. New Trenchstop-based IGBT topologies also improve the VCE_SAT vs. QTOT trade off. And Infineon continues work on wide bandgap devices as GaN and SiC technologies are still far from the theoretical performance limits for RDS(ON)×A FOM. All in all, the selection of a solid-state power switch involves device qualities, including TCO, system implications, and performance requirements. Optimal electrical performance will come with ground-up designs using wide bandgap devices.
REFERENCES Infineon power, https://www.infineon.com/power Open Compute Project, https://www.opencompute.org Research and Markets global IGBT market forecast, https://www.researchandmarkets.com/reports/4718358/ global-igbt-market-forecasts-from-2018-to-2023
2 • 2020
DESIGN WORLD — EE NETWORK
27
POWER ELECTRONICS HANDBOOK
A typical high-voltage connector for an EV. Real-life use differs dramatically from conditions in derating tests where 1.4-m-long cables extend from either side of the connector.
Hot topic: Electric vehicle components get more sophisticated thermal models DR. MICHAEL LUDWIG, T&C CORE TECHNOLOGIES
The high electrical currents of fast charging are forcing designers to invent standard techniques for predictably modeling heat and cooling.
TE CONNECTIVITY
THE ONLY WAY electric vehicles (EVs) will ever catch on is if consumers can charge them up quickly. Consequently, manufacturers are developing high-power charging (HPC) schemes which would enable users to get 300 km of range in less than 10 minutes. The current generation of dc chargers work with 50 kW and 125 A in majority, but the next generation is expected to involve up to 350 kW. Simultaneously, designers of EVs are trying to boost range by reducing component weight and increasing efficiency. All of these changes will directly impact the design of the vehicle’s energy backbone.
28
DESIGN WORLD — EE NETWORK
2 • 2020
eeworldonline.com | designworldonline.com
EV THERMAL MODELS
But the drive to reduce weight and improve electric drivetrain performance makes thermal issues tougher to manage. Higher electrical currents generate more heat, while lighter and smaller components make heat sinking more problematic. Thermal considerations for future EV charge cables will only grow more stringent. Connectors attached to cables with an effective cross-section of 95 mm2 are carrying loads of up to about 400 A in lab conditions. However, they are not generally rated to be used under these loads. The current-carrying capability is derated to compensate for degradation over a lifetime and provide a safety margin for application-specific characteristics that might impact the cooling behavior. Derating factors vary regionally between 0.8 and 0.9. Aluminum wires are appealing to automakers as they are beneficial in regard to weight and cost. However, they should be used for smaller diameter cables and wiring with lower electrical loads as the characteristics of the material and the interconnection to aluminum needs to be carefully thought-out for thermal considerations when joule heating becomes dominant phenomenon. These complications certainly affect terminals and connectors used to handle EV
power. Traditionally, terminals and connectors have power ratings derived from derating measurements such that the terminals and connectors operate at less than their maximum capabilities to prolong operating life. This practice has led to robust designs that have significant safety margins. But this practice doesn’t consider common applications in which the plug mounts to an aggregate; the interactions with busbars; and the transient behavior of components. TE Connectivity is collaborating with ZVEI, a leading German association for electrical and electronic manufacturers, to develop a framework that overcomes these limitations. The framework will enable effective evaluation of how wires and connectors perform when they are integrated in relevant applications, while also considering their specialized boundary conditions. Today’s testing models have limitations. Typical applications for high-voltage (HV) connectors in EVs involve mounting one side of the connector to a busbar, such as at the battery, the inverter, or the onboard charger. This application differs strongly from the derating test scenario in which cables 1.4 m long are applied to both sides of the connector. In addition to boundary conditions,
the ambient environment plays a dominant role in many applications. Cooling strategies employing forced convection can boost current carrying capability, whereas cable ducts hinder convection. The derating scenario has limited usefulness because it considers only free convection when deriving HV connector specifications. In the characterization of current-carrying capabilities, conditions based on derating consider only steady-state temperatures. However, the transient behavior of components is actually more important. To illustrate, when a current of 400 A is applied to a connector through 1.4-m of cable, it takes more than an hour for the assembly to reach a steady state. In real life, a high-power charging solution will complete the charging operation within 10 minutes. So charging finishes long before components reach a steadystate temperature. In addition, real driving cycles often are characterized by highly dynamic load changes. These changes produce high-amplitude current peaks that last for a few seconds. The need to accommodate these peaks is an applicationspecific requirement that doesn’t show up in conventional specification methods. The new approach to evaluating HV
Typical EV electrical systems
Climate compressor
Heater
A block diagram of the typical electrical system for an electric vehicle.
Motor
Onboard charger (AC/DC)
Charging inlet (AC/DC)
HV battery system, including: • • • Power distribution
AC DC Inverter
DC
DC
Motor
Cell connectivity Module connectivity Battery management system (BMS) DC to 12 volt - further system not displayed
AC Inverter
Converter
eeworldonline.com | designworldonline.com
2 • 2020
DESIGN WORLD — EE NETWORK
29
POWER ELECTRONICS HANDBOOK Abrupt load changes characterize driving cycles
50
30
0
20 10 0
0
50
100 time / sec
connector performance uses thermal equivalent circuits (TEC) which are analogous to electrical equivalent circuits. These TECs consist of thermal resistances, the equivalent to ohmic resistances, and thermal capacities, the equivalent to capacitors. Heat is generated at various positions within the TEC according to the law of Joule heating. Additionally, the modeled part interacts with the ambient environment via radiation and convection at the surface. A third and dominant heat transfer mechanism is heat conduction towards the aggregate system and individual cables. Values for the thermal resistors can be derived from geometrical considerations and specific thermal resistance, while the capacitor values can be estimated based on specific heat capacity and weight. These considerations are valid for primary heat transport along the current-carrying path. However, another approach is needed to derive parameter values for heat transport mechanisms involving isolation and shielding. Consequently, there is a second approach that relies on data generated by finite element methods (FEM) and empirical data generated on a dedicated test bench.
A HARMONIZED SIMULATION MODEL ZVEI is defining harmonized TECs for connectors and cables. As a result, products are characterized not only by derating curves but also by a parameter set describing the component behavior. Engineers can use component TECs and parameter sets in a
150
Aggregate Connector header
Connector plug
Aggregate Aggregate
30
DESIGN WORLD — EE NETWORK
2 • 2020
A graph from the worldwide harmonized light vehicles test procedure (WLTP) driving cycle reveals the abrupt load changes characterizing real driving cycles. Here, vehicle speed is in blue and normalized power dissipation is in red.
200
modular fashion to build thermal models and customize then to specific application conditions. A harmonized simulation model helps find the best thermal solution and analyze the effects of various influencing factors. Influencing factors, which can be adjusted easily, typically include current profiles linked to charging or driving cycle data, parameters which underly aging effects, effects tied to wire length or cross-section, application of a radiator to improve convection, and other variables. A non-obvious advantage of this method is that it makes it possible to measure the temperature of inaccessible components. Parts with low mass react immediately to thermal stresses before conduction can cool them off. Measurement of physically inaccessible components lets manufacturers better understand overall temperature generation and distribution and find performance bottlenecks without the need for extensive FEM simulations. The TEC is simple enough that there are no special hardware requirements to make it work. Simulations containing models for an aggregate, a harness, and a connector can run within a few seconds, simulating a transient current profile over a few hours with a resolution of seconds or even less. From a software perspective, any SPICE simulation environment can be used as well as any physical modeling software. In a nutshell, the evolution of EVs will require new approaches
Thermal equivalent circuit for EV connections Aggregate
power / a.u.
power / a.u.
40
Harmonized TECs for EV connectors and cables are simple. Simulations containing models for an aggregate, a harness and a connector can run within a few seconds.
eeworldonline.com
|
designworldonline.com
EV THERMAL MODELS Real part geometry
Thermal equivalent circuit Heat source Current I/O
An example of a thermal equivalent circuit (TEC) and the part it represents. Here, thermal resistances are analogous to ohmic resistances and thermal capacities are analogous to capacitors. The mathematics of both kinds of circuits are the same.
to developing and testing high-power connectors which cannot be addressed using conventional specification methods. A methodology based on thermal equivalent circuits helps designers assess transient behavior stemming from highly dynamic driving profiles or short high-power charging applications. TE Connectivity is collaborating with ZVEI to develop a general guideline to standardize models for both connectors and for shielded and unshielded cables which will enable more accurate testing and fit-for-purpose engineering. The thermal analysis reflects specific properties of various connectors through use of an intrinsic parameter set applied to the standardized models. Manufacturers can execute simulations using any SPICE or physical simulation software on an office PC or desktop. The ability to do efficient computations also lets teams simulate large systems, including aggregates, connectors, and wires, as well as the thermal response to current profiles over a few hours, as recorded in real driving cycles.
Heat flow I/O
Current I/O
Resistance
Resistance
Heat flow I/O
Capacity Initial temperature
TO
SORBOTHANE CAN PROTECT PRODUCT PERFORMANCE FROM SHOCK AND VIBRATION HARM
PROVIDING
SHOCK, VIBRATION & NOISE
SOLUTIONS
REFERENCES
Visit sorbothane.com for Design Calculators and Technical Data
TE connectivity solutions for hybrid and electric mobility, www.te.com/high-power-charging
MADE IN THE U.S.A.
800.838.3906
sorbothane.com DESIGN WORLD — EE NETWORK
31 Sorbothane_DesignWorldAd_A.indd 1
12/11/19 2:04 PM
POWER ELECTRONICS HANDBOOK
Applying large banks of supercapacitors It pays to know techniques for mitigating leakage current and overvoltages in uses where several supercapacitors work in parallel.
MICHELE KINMAN
ADVANCED LINEAR DEVICES INC.
IN APPLICATIONS REQUIRING rapid charge/ discharge cycles or short-term energy storage, you’ll
Precision dual SAB overt voltage protection PCB SABMBOVP2XX schematic diagram V+
often find supercapacitors connected in series or parallel. However, there’s a potential problem when supercaps are wired together in banks: No two
R X1
supercaps are identical, a fact that may lead to a
R X2 3
slight voltage imbalance between them. More specifically, these imbalances arise because the individual capacitors in capacitor assemblies each have minuscule variances in their make-up that contribute to slight differences in their electrical properties such as capacitance, internal resistance, and leakage current. In particular, supercapacitor leakage current depends on parameters such as aging, the material/construction of the supercapacitor, and the operating bias voltage. Leakage current is also a function of the charging voltage, the charging current, operating temperature range, and the rate-ofchange of many of these parameters. The usual way of accommodating these changing conditions is with a balancing circuit. Here, added balancing circuitry ensures weaker capacitors don’t drain stronger units during discharge and that individual capacitors don’t see overvoltages during charging. Large-cell supercapacitors, in particular, require over-voltage balancing because these cells can incur large energy flows. The simplest supercapacitor balancing circuit consists of a resistor put in parallel with the capacitor terminals. Resistors with the same value in parallel with all cells allow cells with higher voltages to discharge through the external resistor at a higher rate than the cells with lower voltages, thus distributing the total capacitor bank voltage evenly across the capacitors.
32
DESIGN WORLD — EE NETWORK
2 • 2020
8,2
4
ON1 J6
7
1,5
R X4 6
R P1
ON2 J8
VA J2
D1 C1
M1 R P2
R X3
Q1
J1
OP1 J7 VB J3 Q2
D2 C2
M2 R P4
R P3
OP2 J9 VC V-
J5
J4
The SABMBOVP2XX schematic. The circuit can be viewed as a precision voltage clamp that functions like a Zener diode. Typically, the clamp current changes from a few nanoamps to over 100 mA (about 1,000,000 times higher) at the clamp voltage within a 100-mV transition.
eeworldonline.com | designworldonline.com
APPLYING SUPERCAPS
GPS
Automation control unit Houses the vehicle’s on-board intelligence and executes all automation and assistance functions. Collects data from the vehicle’s numerous sensors and combines it to give a comprehensive view of the surrounding area. The control unit also receives transport missions from the offboard logistics system and translates them into instructions that the vehicle systems can understand.
The GPS gives the vehicle’s position down to a few meters and allows the vehicle to plan its route. The maneuvering required to follow the route is supported by the sensors and data fusion.
Mobile data link
The mobile data link is the vehicle’s communication channel for receiving transport missions, reporting its status and performance, and sharing perception data with other autonomous vehicles.
Powertrain system
Scania’s intelligent powertrain handles the truck’s propulsion with the highest precision and energy efficiency. The central powertrain control system controls the engine, gearbox, clutch, and auxiliary brakes.
Short range radar
Mounted at each corner of the vehicle, the short-range radars provide 360-degree detection of other vehicles and pedestrians. They function in all weather and light conditions.
Inertial sensors
Multi-lens camera
The inertial sensors measure the rotation and accelaeration of the vehicle to help the automation control unit calculate how it is moving.
Long range radar
Wheel speed sensors
By measuring the rotation of each wheel, the automation control unit can calculate how the vehicle moves and turns.
The problem with this scheme is that the resistor constantly dissipates power as it conducts. When capacitance values of supercapacitors become larger, for values ranging from 100 F to 1,000 F, the capacitor mismatch problem becomes more pronounced and more difficult to correct, especially if capacitor balancing also requires ultra-low power consumption. The typical approach is to devise a balancing circuit containing a switch (usually a MOSFET) in series with a resistance. The switch keeps the balancing circuit off, and dissipating no power, unless cells are out of balance. One difficulty with this approach is that the small MOSFET current meant for correcting leakage current imbalances may be too small if there is too much variation among the supercap cell tolerances. Designers avoid this problem by characterizing the supercaps they use and ganging together those having capacitive values that are close to each other.
TYPICAL APPLICATIONS Experts say 5G base station infrastructure could consume three times more power than its 4G LTE predecessor. One reason is that 5G needs eeworldonline.com | designworldonline.com
With its range of up to 200 m in front of the vehicle, the long-range radar enables high speed driving.
Automation control unit
EAS is an electrohydraulic system that enables the automation and assistance functions to safely steer the vehicle along roads and around obstacles.
Mounted behind the windscreen, the multi-lens camera monitors the area in front of the vehicle to detect objects, vehicles, pedestrians and lane markings. With stereoscopic vision, it can see the shape of the ground in much the same way a human can.
An autonomous heavy duty vehicle three times as many base as envisioned by Scania in Sweden. stations for the same coverage as LTE due to the higher frequencies used. In addition, typical 4G base stations now use four transmitter and four receiver (4T4R) elements, while 5G is expected to use 64T64R MIMO arrays. Consequently, there is much interest in making 5G base stations super energy efficient. One point to note is that today, a 4G base station consumes a significant amount of energy even when there is no output power. The reason is that most 4G circuitry remains active to transmit mandatory idle mode signals -- such as those for synchronization, reference, and system information --defined in the 4G standard. But 5G NR base stations (i.e. those that are not backwardcompatible with LTE) will be in a sleep state when there is no traffic to serve. NR entails far fewer always-on signaling transmissions than LTE (where required signal transmissions are typically less than 1 msec apart). Thus 5G NR base stations are expected to see both deeper and longer periods of sleep with little or no ongoing data transmissions, significantly enhancing overall network energy consumption. 2 • 2020
DESIGN WORLD — EE NETWORK
33
POWER ELECTRONICS HANDBOOK
5G NR base stations coming out of sleep mode will have spiking power demands that are well suited to power supplies incorporating supercapacitors. Here, use of supercaps can enable use of a power supply that is smaller than would otherwise be necessary, further boosting energy efficiency. Ditto for the UPS used to back up 5G NR base stations. Other examples of applications that need large supercapacitor banks and near-zero power waste include autonomous transport systems ranging from large off-road vehicles to small robotic rovers loading and unloading trucks, as well as automated guided vehicles that roam the factory floor. Supercapacitors, typically with values of 3,000 F at 2.7 V, are linked in series and parallel in these applications. Large trucks and off-road vehicles today still generally have a propulsion system incorporating an internal combustion engine and an intelligent powertrain which needs a high-precision and energy efficient power supply; it can be challenging to design their energy storage systems to handle high-transient-current conditions such as cold starts. (Electric locomotives have similar starter systems.) In this regard, supercapacitors will instantaneously supply pulse power to kick-start these large engines, even from a cold start. This past Dec., General Motors won a 2019 R&D 100 Award
for its capacitor-assisted battery (CAB) technology that it intends to use on future GM vehicles. This hybrid cell increases cold-cranking battery performance and allows the electrical power system to offset temperature impacts on lithium-ion battery supplies. Developed at GM’s China Battery Lab, the new hybrid cell is expected to impact lowvoltage and hybrid electric vehicle systems, as well as non-automotive applications requiring high-power response. It has already been licensed to two global battery manufacturers for mass production. In autonomous transport systems, supercapacitors must also be able to work over a high duty cycle and handle frequent deep discharges. Supercapacitors help power start/stop systems as well as help regenerative braking systems work better. Railways are exceptionally energy efficient in terms of energy per passenger mile. (The American Bus Association puts energy use of light rail at about 1,800 btu/passenger mile. For comparison, ABA estimates the average car trip at about 3,800 btu/passenger mile.) Rail cars use regenerative braking to absorb energy and store it in both batteries and supercapacitors. Supercapacitors are perfect in this application. Besides sitting on the train itself, banks of supercaps can be installed at stationary points along the track. Consider traction power substations, which provide power to the
Proven integrity AND industry know-how Electrocube is one of the most respected design manufacturers of passive electrical component products for a wide range of standard and custom applications – from aerospace and audio to elevators and heavy equipment – as a capacitor supplier, resistor-capacitor distributor, and more.
Bishop Electronics, Seacor, Southern Electronics, F-Dyne
34
DESIGN WORLD — E EELNETWORK E C T R O C U B E . C O M |2 •82020 0 0 . 5 1 5 . 1 1 1 2 | I N F O @ E L E C T R O C Ueeworldonline.com BE.COM
|
designworldonline.com
APPLYING SUPERCAPS Light-rail electrical system Train (t) Technical electrical drive
A typical dc light railway electrification system. This is a single train which is traveling between two substations.
Motors
Running rail (rr) Third rail (tr)
GAP
GAP
Contact shoe
Substation (ss) (Transformer rectifier)
overhead catenary system or third rail. The substation converts utility power from 13.2 kVac to the 750 Vdc that the trains use and transmits it via the contact system. Energy generated by decelerating vehicles typically feeds into the traction electrification system incorporating supercapacitors. Supercaps usually sit in substations to power rail cars. The supercaps get energy both from the electrical utility and fed from rail cars themselves as their regenerative braking systems decelerate them. Alternatively, some rail installations also use capacitor-equipped rail cars to absorb regen power. Placing supercapacitor units onboard rail cars reduces power substation loads and permits the light-rail system to expand without the expense of additional substations. One reason supercapacitors make sense for light rail is that rail cars might go through 100,000 to 300,000 cycles of acceleration, running, and deceleration annually. Many battery technologies would wear out quickly with this sort of load profile, but it is no problem for supercaps. Additionally, supercapacitors are lighter than a comparable battery, advantageous when the supercap sits on a rail car. Supercapacitor banks also have a place in low-inertia power grids. Here grid inertia refers to the kinetic energy in the electricity grid. This kinetic energy is that of the conventional eeworldonline.com
|
designworldonline.com
Substation (ss) (Transformer rectifier)
generators at power plants which rotate at the same frequency as the electricity grid. Thus low-inertia systems are those with a significant amount of non-synchronous renewable energy sources such as solar and wind. The problem with low-inertia grids is they may not be able to provide the inertia to support grid frequency and can be susceptible to issues of power quality and blackouts. When power demands spike, the frequency of the grid tends to drop. In grids having a great deal of traditional rotating generator capacity, the rotating mass functions as a shock absorber in the event of sudden heavy loads. Solar panels, of course, have no rotating mass. And because wind turbines connect to the grid through a frequency converter, the wind turbine’s rotating mass doesn’t provide inertia when grid frequency drops. Supercapacitors can play a role in grid stabilization by providing reactive power, power attributed to ac current and voltage that is out-of-phase with each other. Reactive power, measured in volt-amperes-reactive, helps regulate grid operation and is necessary for operating loads such as motors and transformers. Reactive power does not travel as far as real power. Long transmission lines operating at heavy loads can cause conductor heating and falling voltages because of the reactive power component associated with the 2 • 2020
distributed L-C of the power line. In this regard, supercaps can be positioned close to power loads to serve as a fast source of reactive current. Lowering the reactive current demand allows the delivery system to carry more real current and helps the utility maintain its service voltage within required limits. Supercapacitors can as well play a role in renewable energy applications. Applications powered by solar panels typically incorporate deep-cycle batteries sized for the application’s total watt-hours per day of use. Supercapacitors used as an addition to the battery capacity brings several benefits. Perhaps most obvious is that supercaps can handle short-term spike loads to permit use of a smaller and less costly battery bank. The fact that supercapacitors use physical rather than electrochemical charge storage and so have a practically unlimited cycle life comes in handy as well. And because supercap equivalent series resistance (ESR) doesn’t vary appreciably with temperature, supercaps can sit outdoors near the solar array. In the same vein, supercaps can serve as a backup source of dc power for data centers. The vast majority of power glitches (87%) last less than a second. A supercop UPS system can sit in a smaller footprint and provide a higher power density than a battery backed-up UPS. DESIGN WORLD — EE NETWORK
35
POWER ELECTRONICS HANDBOOK Example: Supercap power bufffer in energy harvesting Supercapacitor with high C (energy) and low ESR (power) buffers the source from the load. Source sees average power. Load sees low impedance source.
Interface electronics
Energy harvesting source must be able to deliver average load current = 420μA
Using a supercapacitor as a power buffer.
DC:DC converter can be sized for average load power
Load. Hi power peaks
e.g. 60% efficient. i/p power= 0.75mW/60% = 1.25mW. Use a charge pump to charge supercapacitor. Input current ~ 420μA.
Report status 1/hr with GPRS class 8; 3 sec SMS: 2A @ 3.6V/8 x 3/3600 = 0.75mW average power, with 7.2W peak power
UPS electronics can typically work up to 3540°C without derating or a degradation in performance. But a UPS with sealed lead-acid (SLA) batteries has an optimum range of 2025°C – the recommended temperature range for an SLA battery to reach its design life. Finally, a supercap UPS recharges almost instantaneously compared to the time necessary for recharging a traditional battery set through the UPS rectifier or external battery charger. These are only a small sampling of the applications where supercapacitors contribute to a reliable system design with a long lifespan. One development that helps promote reliable supercap applications comes from Advanced Linear Devices which has a precision dual-channel PCB called the SABMBOVP2XX family that not only automatically balances supercapacitors but has an additional over-voltage protection (OVP) feature. One more added benefit of this series is that it performs all these tasks at ultra-low power with leakage current regulation. It can be said that this solution has zero power burn.
REFERENCES SABMBOVP2XX family, www.aldinc.com/pdf/SABMBOVP.pdf
36
DESIGN WORLD — EE NETWORK
2 • 2020
eeworldonline.com
|
designworldonline.com
You probably already use Tadiran batteries, but just don’t know it!
PROVEN
40 YEAR OPERATING
LIFE
If you have a smart automatic water, gas, electricity, or heat meter in your home. If you have an electronic toll collection transponder, tire inflation sensor, or emergency E-CALL system in your car. If you have a GPS tracking device on your trailer, container, or cargo. If you have wireless sensors, controls, or monitors in your factories and plants. If you use electronics with real-time clock or memory back-up in your office.
If you have never heard of Tadiran Batteries, it is only because you have never had a problem with our products powering your products. Take no chances. Take Tadiran batteries that last a lifetime.
* Tadiran LiSOCL2 batteries feature the lowest annual self-discharge rate of any competitive battery, less than 1% per year, enabling these batteries to operate over 40 years depending on device operating usage. However, this is not an expressed or implied warranty, as each application differs in terms of annual energy consumption and/or operating environment.
Tadiran Batteries 2001 Marcus Ave. Suite 125E Lake Success, NY 11042 1-800-537-1368 516-621-4980 www.tadiranbat.com
*
POWER ELECTRONICS HANDBOOK
The try-before-you-buy route to energy eff icient power design XUNING ZHANG, LEVI GANT
LITTELFUSE, INC.
Evaluation platforms help pick SiC components in power conversion circuits. THE QUEST FOR greener energy production and consumption has put a premium on high-efficiency power circuitry. In that regard, many modern power supplies and converters operate at much higher voltages that allow use of lower currents to minimize I2R losses. Silicon carbide (SiC) MOSFETs and diodes are important elements of these new high-power, high-voltage power conversion circuits. SiC MOSFETs provide low on-resistance and can switch back and forth between on and off states rapidly. Consequently, they dissipate much less power than insulated gate bipolar transistors (IGBTs) which have slower turn-off speeds and higher turn-off switching power loss. In addition, silicon carbide’s wide bandgap enables SiC devices to operate at high voltages. In contrast, silicon-based MOSFETS can’t realize both high blocking voltages and low on-resistances. As a result, SiC devices are becoming integral in high-power applications. Because of the high power levels SiC devices deal with, designers must evaluate both the SiC devices themselves and their gate driver circuits. SiC technology is still relatively new, and device performance under a wide range of conditions is not fully characterized. An evaluation platform will enable design engineers to evaluate SiC MOSFETS, SiC Schottky diodes, and gate driver circuits under continuous operation in converter circuit applications. The evaluation platform will aid in accelerating design cycles for successful, SiCbased power converter design and assist in speeding the time-tomarket for the end product.
38
DESIGN WORLD — EE NETWORK
2 • 2020
The gate drive evaluation platform includes the motherboard, two plug-in gate driver modules, and the heatsink and fan to support up to 5 kW of output power.
eeworldonline.com
|
designworldonline.com
SiC EVAL BOARDS Typical gate driver evaluation setup DC +
PWM & protection
Gate driver board
+12 V
PWM & protection
OUT Gate driver board DC -
DESIGN CHALLENGES FOR POWER CONVERSION CIRCUITS To maximize power output and efficiency of power conversion circuits, designers must ensure that: •
• • •
• •
•
The power devices can perform at rated power and current and deliver enough power to the load The circuit minimizes internal power loss for maximum efficiency The design incorporates protection circuitry for the SiC power devices Printed circuit board (PCB) layout minimizes parasitic inductances and capacitances EMI emissions are within allowable limits The design uses a minimum of passive components to help keep down cost, size, and weight The gate driver helps realize the above goals and assists in maintaining thermal performance within specified temperature ratings.
A gate drive evaluation platform helps designers address all these challenges. The platform can operate at high power levels continuously to characterize the performance of the selected SiC MOSFETS and diodes. The platform also enables the comparison of different gate drivers under multiple test conditions. Gate drivers can be evaluated for thermal performance, EMI immunity, and the capability to drive the power components so they operate at high efficiency. Finally, eeworldonline.com | designworldonline.com
Simplified diagram of the gate drive evaluation platform. The power configuration is a half-bridge output stage. Not shown are the decoupling capacitors positioned close to the SiC devices to maintain the supply voltage during device switching. The decoupling capacitors and the capacitor across the SiC devices act as a low-pass filter to remove switching noise on the dc supply line.
the platform allows analysis of the design for efficiency improvements, electromagnetic interference (EMI) emissions, cost, size, and weight. The gate drive evaluation platform is essentially a power stage reference design consisting of a motherboard with two SiC MOSFET-SiC Schottky diode pairs in a half-bridge configuration. The half-bridge circuit can output a maximum of 5 kW with an 800-Vdc bus voltage. The motherboard can accommodate two separate gate driver module boards, one for each switch position. Thus, different gate driver integrated circuits and gate driver designs can quickly and easily mount on the motherboard to evaluate gate driver performance and how the driver impacts output power. The third major element of the gate drive evaluation platform is the thermal management, a heatsink and a fan which cools the MOSFET-diode pairs. The heatsinkfan subsystem enables the power circuit to deliver up to 5 kW continuously with the MOSFET-diode pairs switching at frequencies up to 200 kHz. The gate drive evaluation platform’s printed circuit board layout minimizes both loop inductance and coupling between the power circuit and the gate circuit. The two gate driver circuits allow independent evaluation of both the top and bottom gate-driving qualities. The selection of SiC MOSFETs and diodes and the selection of the gate driver are the most important decisions for the power conversion design. The MOSFET must have
2 • 2020
the voltage, current, and power specifications to meet the converter requirements. The gate driver has more sophisticated requirements. It should have a wide voltage range and enough output current to drive the power MOSFET. Recommended drive voltages are 15 to 20 V to switch the MOSFET to its onstate and a voltage of 0 to -5 V to switch the MOSFET to the off-state. The peak output current for the gate drive can range from 1 to 15 A depending on the MOSFET power handling capacity. The driver needs to provide a high pulse current to reduce MOSFET switching loss during switching transients. In addition, high continuous current with small external gate resistances
Parasitics in the gate drive loop VCC Cgd_stray OUT
Rext
Lstray Cgd_stray
VEE Driver IC
Gate drive loop parasitic capacitance and inductance in the gate drive loop.
DESIGN WORLD — EE NETWORK
39
POWER ELECTRONICS HANDBOOK
Waveforms resulting from test conditions: input voltage = 800 V, output voltage = 400 V, switching frequency = 100 kHz, and output power = 2.5 kW.
40
DESIGN WORLD — EE NETWORK
reduces driver temperature during high frequency switching of the SiC MOSFET. The high dv/dt caused by fast SiC MOSFET switching makes high commonmode currents flow through the gate driver and the rest of the power conversion circuit. High common-mode currents can affect the voltage reference node in the control circuit, causing incorrect operation. The magnitude of the common-mode current is determined by the MOSFET dv/dt and the impedance in the common-mode current path. Consequently, the gate driver IC and its power supply both need a high isolation impedance to reduce common-mode current. The isolation capacitance of the gate driver should be less than 1 pF. The isolation capacitance of the power supply should be under 10 pF. Traditionally, optocouplers would provide the isolation. Newer IC technology can employ inductive or capacitive
2 • 2020
isolation. The new methods are known as digital isolator techniques. The optocoupler and the digital isolator have both advantages and disadvantages. The optocoupler sources current which makes its input less susceptible to EMI. However, optocouplers can’t handle data transmission rates as high as those of digital isolators and bring longer pulse-width distortion times. Pulse-width distortion time refers to signal delay time through the driver IC. In a half-bridge power conversion topology, excessive delay can create waveform distortion and low frequency noise. Optocoupler performance varies with the drive voltage, temperature, and device age. Digital isolator-based drivers have more stable parameters over temperature. Because digital isolators operate with a voltage input, they can be more susceptible to EMI. But all in all, the digital
eeworldonline.com
|
designworldonline.com
SiC EVAL BOARDS Comparing switching transients (turn-on), IXDN614 and IXDN602
(a) 10 Ohm gate resistor
MOSFET turn-on transient with two different driver ICs and a 10-Ω gate resistor. isolator’s more stable operating parameters makes it a better choice than optocouplers in gate drivers for power conversion circuits using SiC MOSFETs. With high-power circuits, protection mechanisms are necessary to prevent device thermal runaway and device and circuit damage from fault conditions. Gate driver ICs that incorporate protection circuitry are highly recommended. Gate drive ICs should have de-saturation (de-sat) protection, soft turn-off during a fault condition, a Miller clamp circuit, and under-voltage lock out (UVLO). eeworldonline.com
|
designworldonline.com
(b) 1 Ohm gate resistor
MOSFET turn-on transient with two different driver ICs and a 1-Ω gate resistor. De-sat protection circuitry turns off a MOSFET in the event of a load short-circuit. Soft turn-off avoids a large transient voltage overshoot and turns off the MOSFET during a shoot-through failure (where both MOSFETs are momentarily on simultaneously). A Miller clamp circuit prevents the shoot-through condition by draining current from the parasitic drain-gate capacitance to avoid a transient rise in the gate voltage. The clamp circuit keeps the MOSFET from turning-on when it should be in the off-state. If voltage supply for either the gate driver input or for the isolated output circuit gets too low, the UVLO circuit 2 • 2020
DESIGN WORLD — EE NETWORK
41
POWER ELECTRONICS HANDBOOK Evaluating switching losses +12 V power supply
Digital controller or function generator
DC +
Gate driver board
+12 V power supply
OUT External output filter
Gate driver board
Gate driver evaluation platform
Gate driver switching loss test using a buck converter as a load. Visible here are the gate drive voltage, MOSFET drain-source current, and MOSFET drain-source voltage.
turns off the gate drive to protect the MOSFETs from improper switch timing. These protection circuits ensure a more robust and safe power conversion circuit. PCB board layout has a major impact on the performance of dynamic circuits such as high-efficiency power conversion circuits. Parasitic capacitance and inductance from PCB traces and ground planes add to the parasitic capacitance and inductance in the circuit. Parasitic components in the gate drive loop degrade MOSFET switching performance. Gate-source capacitance forces a higher driving current from the gate driver IC. Stray inductance boosts gate-source voltage overshoot and leads to ringing during MOSFET switching. To minimize the stray capacitance and inductance, keep the gate path as short as possible by placing the gate driver, the gate resistor, and the decoupling capacitor close to the MOSFET gate. Minimize loop inductance by routing the gate return path directly below the gate supply trace. Maximize the distance between the MOSFET gate traces and the drain traces to reduce the size of the gate-drain capacitance. This practice cuts the current entering the gate which reduces the Miller effect. Additionally, ground planes under power conversion circuits add capacitive coupling; avoid use of ground planes for MOSFET switching-based power conversion circuits. All these PCB layout recommendations have been implemented in the gate drive evaluation platform to avoid design, layout, and test of a custom test board. The gate drive evaluation platform can easily compare switching loss and switching transients using different gate drive ICs. Consider the case of evaluating gate drivers for a buck converter operating under continuous switching conditions. The buck converter operates at 100 kHz and will output 2.5 kW. The drive capabilities of driver ICs and the external gate resistances used will influence the SiC MOSFET switching transients and the overall switching losses. In this test, the first gate driver has a drive current rating of 14 A and the second has a drive current rating of 2 A. Each gate driver was tested with a 10-Ω and a 1-Ω gate resistor.
42
DESIGN WORLD — EE NETWORK
DC link capacitor
2 • 2020
DC -
Load resistor
Probes
DC source
Oscilloscope
The 10-Ω gate resistor eliminates differences in the performance of the gate drivers. The 10-Ω gate resistor does slow MOSFET transient switching speed which increases switching loss. The differences between the high-output-current driver and the low-output-current driver are more significant. The MOSFETs switch faster when the high output current driver is used with a lower gate resistance. The lower gate resistance does exhibit more ringing during a switch transition than the higher gate resistance. The designer must find the optimum combination of gate driver, gate resistance, and MOSFET to minimize switching loss. The gate drive evaluation platform can help evaluate driver IC thermal performance thanks to the heatsink and fan that enable the MOSFETs to operate in a continuous switching output state. The platform can also be used to test driver protection functions. In a nutshell, gate drive evaluation platforms are a tool that facilitates the evaluation of SiC devices and the gate drivers. With gate drive modules plugged into the motherboard, designers can easily compare efficiency and thermal performance with different gate driver ICs. Designers can use the PCB layout techniques on the evaluation platform and the component recommendations to overcome the design challenges of SiC devices to develop efficient, thermallycontrolled, and protected power conversion circuits. As a result, the evaluation platform enables faster design of efficient power conversion circuits and speeds product time-to-market.
REFERENCES Littelfuse application note on the gate drive evaluation platform, littelfuse.com/gdevappnote
eeworldonline.com | designworldonline.com
A lot to see for PoE
Coilcraft off-the-shelf PoE transformers for high-power IEEE 802.3bt applications.
Coilcraft has off-the-shelf power transformers for all your high-power PoE applications, up to 71 W (input) at PD! Coilcraft offers PoE transformers for a variety of power levels compatible with IEEE 802.3af/at, as well as the new high power IEEE 802.3bt standard (up to 71W PD input power). Available in Flyback and Forward-mode models, our PoE transformers provide excellent power conversion efficiency and
high isolation voltages within the smallest package sizes possible. They also support a wide variety of standard output voltages to suit a broad array of PoE powered devices. Learn more about our off-the-shelf PoE power transformers and order free evaluation samples at www.coilcraft.com/PoE. ÂŽ
WWW.COILCRAFT.COM
POWER ELECTRONICS HANDBOOK
The 40-year battery pack SOL JACOBS
TADIRAN BATTERIES
Not all primary batteries offer the low annual self-discharge rate needed for the lengthy lifespan of remote wireless devices consuming microamp-level currents. THE VAST MAJORITY OF remote wireless devices are powered by primary (non-rechargeable) batteries that range from consumer grade to industrial grade. Consumer-grade alkaline batteries deliver higher discharge rates of energy, resulting in a short operating life. Also available are consumer-grade lithium batteries that deliver medium-to-high discharge rates of energy with short-tomedium operating life, including iron disulfate (LiFeS2), and lithium manganese dioxide (LiMNO2) chemistries. Meanwhile, a growing number of industrial wireless sensors and devices are being designed to operate for decades without having to replace the battery. They typically require only a low rate of energy discharge and consume microamps of average current. Many of these devices sit in extreme environments and hard-to-access locations, often connected to the Industrial Internet of Things (IIoT).
44
DESIGN WORLD — EE NETWORK
2 • 2020
Lithium-based batteries have high intrinsic negative potential, exceeding that of all other metals, with an operating current voltage (OCV) ranging from 2.7 to 3.6 V. Lithium batteries are also non-aqueous, with the absence of water enabling them to endure extreme temperatures without freezing. Among all commercially available available chemistries, bobbintype lithium thionyl chloride (LiSOCl2) cells are preferred for use in remote locations and extreme environments. Bobbin-type LiSOCl2 batteries feature the highest capacity and highest energy density of any lithium chemistry, along with an extremely low annual self-discharge rate (less than 1% per year), enabling certain devices to operate for up to 40 years. Bobbin-type LiSOCl2 chemistry also features the widest possible temperature range (-80 to 125°C), along with a glass-tometal hermetic seal that resists battery leakage. Common applications include AMR/AMI metering, M2M, SCADA, tank-level monitoring, asset tracking, and environmental sensors, to name a few. eeworldonline.com
|
designworldonline.com
40-YEAR BATTERY PACK Medium rate bobbin-type lithium thionyl chloride 10 batte-year ry lif e
Load size
Selfdischarge 0 years
10 years
Low rate / low self-discharge bobbin-type lithium thionyl chloride 40 batte-year ry lif e Selfdischarge 0 years
10 years
20 years
30 years
40 years
High rate / low self-discharge bobbin-type lithium thionyl chloride with hybrid layer capacitor 40 batte-year ry lif e
Load size
The battery marathon analogy: Distance is equivalent to the battery/ device operating life. The energy necessary to run up an incline is equivalent to the rate of battery self-discharge. The higher the self-discharge rate, the steeper the incline. A course with a lot of hills can have runners walking by the end of the race, and higher battery selfdischarge reduces the availability of useful power for device operation and thus reduces the operating life. Additionally, hurdles can be thought of as requiring energy pulses. The higher the hurdle (obstacle) the greater must be the battery’s pulsing ability.
Selfdischarge 0 years
10 years
20 years
30 years
40 years
TLM - lithium metal oxide Load size
20 batte-year ry lif e Selfdischarge
0 years
10 years
20 years
All batteries experience some amount of self-discharge, where cell capacity gets depleted even when the battery is not connected to an external load. The ability to control passivation (basically, reduction of chemical reactivity) helps to lower selfdischarge: a quality unique to bobbin-type LiSOCl2 batteries. Passivation arises when a thin film of lithium chloride (LiCl) forms on the surface of
eeworldonline.com
|
designworldonline.com
the lithium anode, thus impeding the chemical reactions that result in battery self-discharge. When a load is placed on the cell, the passivation layer causes high initial resistance, resulting in a temporary drop in cell voltage until the discharge reaction slowly removes the passivation layer. The process repeats every time the load is removed. Several variables can influence the passivation effect, including: the current 2 • 2020
capacity of the cell, the length of storage, storage temperature, discharge temperature, and prior discharge conditions, as partially discharging a cell and then removing the load can boost the amount of passivation relative to when the cell was new. Passivation can reduce a battery’s self-discharge rate, but too much of it can block energy flow. Different bobbin-type LiSOCl2 batteries can be designed with varying amounts of passivation and self-discharge. These cells can be designed with medium energy flow rates and higher self-discharge (lifespan of 10 years) or with lower flow rates and lower self-discharge to run a marathon (lifespan of up to 40 years). Battery self-discharge rate is also affected by the quality of the raw materials and the method by which the battery is manufactured. For example, a lower-grade bobbin-type LiSOCl2 battery can lose up to 3% of its normal capacity annually due to self-discharge, thus exhausting 30% of its initial capacity every 10 years, making 40-year battery life impossible. DESIGN WORLD — EE NETWORK
45
POWER ELECTRONICS HANDBOOK Battery qualities that include selfdischarge and operating current are sometimes explained in terms of analogies to liquid in variously sized beakers and bottles.
The bottle analogy Comparing battery uses (opening size), flow (discharge), and self-discharge (evaporation)
By contrast, the highest quality bobbin-type LiSOCl2 battery can feature a self-discharge rate of just 0.7% per year, thus retaining 93% of its original capacity after 10 years. A 40-year operating life is feasible. A 40-year operating life is analogous to running a marathon. Distance is equivalent to the battery/device operating life. The energy necessary to run up an incline is equivalent to the rate of battery selfdischarge. The higher the self-discharge rate, the steeper the incline. A course with a lot of hills can have runners walking by the end of the race, and higher battery selfdischarge reduces the availability of useful power for device operation and thus reduces the operating life. Additionally, hurdles can be thought of as requiring energy pulses. The higher the hurdle (obstacle) the greater must be the pulse ability of the battery. Certain applications, such as medical power tools and actuators, require a high rate of energy drain as well as high pulses. These applications draw average current measurable in amps, making them well suited for lithium metal oxide batteries. Flashlights, toys, and other consumer applications that require rates of discharge measurable in milliamps to amps may be best suited for alkaline, LiFeS2 and LiMNO2 batteries that can deliver medium pulses. Many remote wireless sensors and other ultra-long-life, low-drain applications with average current measurable in microamps require the use of standard bobbin-type LiSOCl2 batteries that can run marathons thanks to their extremely low self-discharge rates. But their low-rate design prevents these cells from delivering high pulses without modification. However, standard bobbin-type LiSOCl2 batteries can be easily modified by adding a patented hybrid layer capacitor (HLC). The standard bobbin-type LiSOCl2 cell delivers the low daily background current for a 40-year marathon while the HLC delivers periodic high pulses to power two-way wireless communications (steeple jumping). The patented HLC also features a special end-of-life voltage plateau that can be interpreted to deliver low-battery status alerts. Consumer applications often use supercapacitors that deliver high pulses electrostatically rather than chemically. Supercapacitors are only
46
DESIGN WORLD — EE NETWORK
2 • 2020
selectively used in industrial applications due to performance limitations such as short-duration power, linear discharge qualities that prevent use of all the available energy, low capacity, low energy density, and high annual self-discharge rates (up to 60% per year). Supercapacitors linked in series also require the use of cell-balancing circuits, which adds expense and bulk and boosts their self-discharge rate. eeworldonline.com
|
designworldonline.com
40-YEAR BATTERY PACK TLM - lithium metal oxide
15 A
MP
1 AMP
TLM
Other AA sized high rate cells
SHORT-TERM TESTS DON’T SIMULATE LONG-TERM PERFORMANCE Long-term battery performance cannot be easily duplicated using short-term test procedures, so specialized test methods must be employed. Here are some proven test methods: Long-term laboratory testing – The ideal way to monitor battery self-discharge is to continually test batteries over time under various conditions, covering almost every possible scenario. Over time, the accumulated data points can be used to accurately predict performance based on cell size, temperature, load size, etc. Accelerated testing - The Arrhenius equation--which is based on a two-fold increase of reaction rate for every 10°C rise in temperature--is often used to simulate long-term battery operation. Arrhenius tests are run at 72°C, equivalent to about 32 times the theoretical lifetime of battery at 22°C. However, short-term tests using the Arrhenius method tend to show inaccurate results. Calorimeter testing. An extremely accurate test method is to measure the amount of actual heat energy lost using a state-of-the-art microcalorimeter, which can detect energy dissipation as low as 0.1 W. Heat energy is generated three ways: entropy change, often referred to eeworldonline.com
|
designworldonline.com
A hurdling analogy is often invoked for lithium batteries optimized for handling current pulses. For example, TLM Series lithium metal oxide batteries are designed to deliver high voltage (4.0 V), high current pulses, and continuous high power without any delayed response or passivation effect. Ordinary lithium batteries used to provide high-current pulses can run out of juice long before what might be expected using only specifications for average current drain.
as reversible heat; cell over-protection, often referred to as irreversible heat; and chemical reactions, including self-discharge reactions that affect cell capacity, and side reactions that do not affect cell capacity. Calorimeter testing can be especially useful for measuring losses in battery capacity that arise during long-term storage or from operation (including self-discharge), measurable using thermodynamic equations and cell voltage considerations. To ensure accuracy, the batteries must be stabilized for one year prior to testing, as first-year self-discharge tends to be higher than subsequent years. Lithium titration. Lithium titration can be used to measure available cell capacity. The battery is cut open, and then titration is used to dissolve the remaining lithium to determine its volume. The higher the self-discharge rate, the less lithium will be found in in the cell. Lithium titration can be especially useful in special circumstances as when measuring the effects of extreme temperatures and/or prolonged high pulses. Field results – Batteries working in the field offer the best from of validation for test results. Tadiran has its customers provide randomly sampled batteries taken from long-term deployments to verify the real-life impact of long-term exposure to extreme temperatures and to demonstrate how such conditions accelerate battery
self-discharge. For example, batteries deployed in pioneering AMR/AMI devices from the 1980s exhibited significant amounts of unused capacity after 28 years in the field. Another useful indicator of longterm battery performance is the number of Failures In Time (FITs), measurable in billions of device operating hours for devices in the field. Tadiran batteries have FIT rates ranging between 5 and 20 batteries per billion, extremely low compared to the industry average. Every application has unique power requirements, so it is important to determine whether you need a sprinter (high discharge potential); a medium distance runner (moderate to high discharge rate with fairly low selfdischarge); or a marathoner (running with very low discharge for up to 40 years).
2 • 2020
DESIGN WORLD — EE NETWORK
REFERENCES Tadiran Batteries, www.tadiranbat.com
47
AD INDEX POWER ELECTRONICS HANDBOOK | FEBRUARY 2020
Coilcraft ....................................................................... 43
Newark, An Avnet Company .......................................BC
Digi-Key ...........................................................Cover, IFC
Nichicon (America) Corporation .................................. 40
Electrocube, Inc. .......................................................... 34
Pico Electronics ........................................................... 13
Keystone Electronics Corp. ....................................1, IBC
RECOM Power............................................................... 3
KOA Speer .................................................................. 23
Sorbothane .................................................................. 31
Master Bond ................................................................ 27
Tadiran ......................................................................... 37
SALES Jami Brownlee jbrownlee@wtwhmedia.com 224.760.1055 Mike Caruso mcaruso@wtwhmedia.com 469.855.7344 Bill Crowley bcrowley@wtwhmedia.com 610.420.2433 Jim Dempsey jdempsey@wtwhmedia.com 216.387.1916 Michael Ference mference@wtwhmedia.com 408.769.1188
LEADERSHIP TEAM David Geltman dgeltman@wtwhmedia.com 516.510.6514 @wtwh_david Neel Gleason ngleason@wtwhmedia.com 312.882.9867 @wtwh_ngleason Jim Powers jpowers@wtwhmedia.com 312.925.7793
Publisher Mike Emich memich@wtwhmedia.com 508.446.1823 @wtwh_memich Managing Director Scott McCafferty smccafferty@wtwhmedia.com 310.279.3844 @SMMcCafferty
@jpowers_media
EVP Marshall Matheson
Courtney Nagle cseel@wtwhmedia.com
mmatheson@wtwhmedia.com 805.895.3609 @mmatheson
440.523.1685
@mrference Mike Francesconi mfrancesconi@wtwhmedia.com
2014 Winner
48
DESIGN WORLD — EE NETWORK
2011 - 2019
2 • 2020
W
OR
LD
630.488.9029
@DESIGN eeworldonline.com
|
designworldonline.com
™
dESIgN SOLuTIONS
Design Solutions at your fingertips! Contact us for over 70 years of “in-house” design application and engineering expertise in:
• Stamping • machining • aSSembly • cnc • injection molding Our team is just an email or phone call away. It’s what’s on the InsIde that counts ® E L E C T R O N I C S
C O R P.
View our Dynamic Catalog M70 at www.keyelco.com (516) 328-7500
(800) 221-5510
Request a copy of our Product Design Guide M70
DW THiNK Custom_2-20.indd 1
1/10/20 11:20 AM
1.800.463.9275
newark.com
Discover Over a Million
Engineering Products from Suppliers You Know and Trust 900+ new products each week
Custom services such as kitting, panel meters, enclosures, and many more! Market-leading online community of over 600,000 engineers
