E VK Com Englishmagazines 152

Download as pdf or txt
Download as pdf or txt
You are on page 1of 52

16th Annual Power Supplement

Starts After Page 16

AUGUST 2017

SMART VEHICLES LAUNCH


THE AUTONOMOUS MOONSHOT
Full autonomy may be years away, but the building blocks
to help designers get there are developing rapidly page 14

Also in this issue


How GaN-on-Si high-power
chip technology is changing
luminaire design p9

How ever-evolving data centers will


reshape SerDes and memory designs p12

Understanding and solving industrial robot


design challenges p17

An AspenCore Publication electronicproducts.com


REDEXPERT

REDEXPERT. Würth Elektronik‘s online platform for simple #REDEXPERT


component selection and performance simulation.

 The world’s most accurate AC loss model  Available in seven languages


 Filter settings for over 20 electrical and  Online platform based on measured
mechanical parameters values
 Inductor simulation and selection for  No login required
DC/DC converters  Order free samples directly
 Ability to compare inductance/current and  Direct access to product datasheets
temperature rise/DC current using interactive
measurement curves www.we-online.com/redexpert
2 Contents Vol. 60, No. 2 August 2017

7 15
EDITORIAL STAFF

Greg Lupion . . . . . . . . . . . . . . . . . . Managing Editor


516-667-2379 • [email protected]

Paul O’Shea . . . . . . . . . . . . . . Senior Technical Editor


941-359-8684 • [email protected]

Richard Quinnell . . . . . . . . . . . . . . . . Editor-in-chief,


[email protected] • electronicproducts.com

Majeed Ahmad Kamran ..... Contributing Editor


FEATURES
Patrick Mannion Contributing Editor
9
...............
Solid-State Lighting
How GaN-on-Si high-power chip technology is changing luminaire design Lori O’Toole . . . . . . . . . . . . . . . . . . . . . . . . Chief Copy Editor

12 Memory ICs Nicole DiGiose . . . . . . . . Technical Content Manager


How ever-evolving data centers will reshape SerDes and memory designs Max Teodorescu . . . . . . . . .Digital Content Manager

SPECIALS Jennifer Korszun ................... Assistant Editor

14 COVER STORY Smart Vehicles Giulia Fini . . . . . . . . . . . . . . . . . . . . . . . . . . . Graphic Designer


Smart vehicles launch the autonomous moonshot
Giulia Fini . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cover Design

17 Robotics Lauren Heller . . . . . . . . . . . . . . . . . . Production Designer


Understanding and solving industrial robot design challenges
Subscriber Service 1-866-813-3752
TOUCH POINTS
Subscriber Service Fax 1-847-564-9453

4 Viewpoint: Is the data center the next iPhone for the chip business? Reprints ( Wright’s Media) 1-877-652-5295

5 Publisher’s Perspective: Deeply reported news delivers unique insight


Published by AspenCore

7 Outlook (Technology News): 1225 Franklin Avenue, Suite 400


High-density, self-healing silicon could succeed Li-ion technology Garden City, New York 11530
TEL: 516-667-2300 • FAX: 516-667-2301
20 Product Roundup: Packaging, Cabinets & Enclosures
Victor Alejandro Gao

16th Annual
22 Product Trends: General Manager

Power
Sensors & Transducers
Steve Cholas
Supplement
Starts After 26 New Products: Group Publisher Electronics Group

Page 16 26 Packaging & Interconnections


28 Power Sources
Electronic Products Magazine (USPS

28
539490) (ISSN 0013-4953)—Published
Product Mart monthly by AspenCore, 1225 Franklin Ave-
nue, Suite 400, Garden City, NY 11530. Periodicals postage paid Gar-
den City, NY and additional mailing offices. Electronic Products is
distributed at no charge to qualified persons actively engaged in the
authorization, recommendation or specification of electronic compo-
nents, instruments, materials, systems and subsystems. The publisher
reserves the right to reject any subscription on the basis of informa-
ONLINE MUST-READS tion submitted in order to comply with audit regulations. Paid sub-
SCAN TO DOWNLOAD scriptions available: U.S. subscriber rate $65 per year, 2 years $110.
EP CALCULATOR APP: Single issue, $6.00. Information contained herein is subject to change
Special Report: City-wide savings as simple as IoT without notice. No responsibility is assumed by the publisher for its
accuracy or completeness.
garbage http://bit.ly/2uEn0MP
Postmaster: Send address changes to Electronic Products,
PO Box 3012, Northbrook, Il 60065-3012. Phone:
Embedding AI into tiny devices opens a world of 847-559-7317
possibilities http://bit.ly/2vevSFM ©2017 by AspenCore. ALL RIGHTS RESERVED

Publications Mail Agreement Number


XMOS and Qualcomm launch voice-processing dev kits 40012807.
for cloud-based speech recognition http://bit.ly/2veptdK Return Undeliverable Canadian Addresses to:
Station A PO Box 12, Windsor, ON N9A 6J5

AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS COVER IMAGE: SHUTTERSTOCK


4 Viewpoint

Is the data center the next


iPhone for the chip business?

I
n the late 2000s, when the PC in- environments. And that will reshape
dustry was showing signs of slowing design areas such as SerDes, Gigabit
down, the iPhone erupted onto Ethernet, and interconnects.
the mobile scene and reinvented the Next, memory architectures are
smartphone industry altogether. And being redefined to meet data centers’
the smartphone became a key driver for never-ending demand for speed and
the semiconductor industry at a time of bandwidth. The DDR5 standard is
lumbering PC market growth. around the corner and is going to double
The smartphone party went on for the bandwidth and density over DDR4.
nearly a decade before it also began Moreover, DDR5 promises to deliver
showing signs of a slowdown. Then one improved channel efficiency.
day, Intel announced that it was buy- Also worthwhile to mention is the
ing the FPGA powerhouse Altera. On high bandwidth memory (HBM) version
December 28, 2015, almost everybody of DRAM architecture, which is signifi-
had been wondering about why Intel cantly raising the bar on memory capac-
acquired Altera. ity and data throughput by providing a
Ironically, at that time, the PC hard- wider interface to SoC designs. A greater
ware king was mired in a battle for the memory bandwidth and density closer to
riches of the smartphone, to which it the CPU also brings significant efficien-
had been dreadfully late. What industry cy benefits to server and networking
watchers eventually found out can be systems in data centers.
summed up in two words: data center. The 3D XPoint non-volatile memory
The data center had been rapidly evolv- technology jointly developed by Intel
ing in an increasingly connected world and Micron is another example of how
while catering to new data sources such greater bandwidth and storage capacity
as smartphones, tablets, home appliances, are pushing the limits in the memory
wearable devices, and connected vehicles. realm. Memory chips built around the
Intel already had skin in the data center 3D XPoint technology are expected to
game with its Xeon CPUs, and the world’s arrive later this year.
largest chipmaker wanted to make its bets The GPU and FPGA chips represent
in the data center business even safer. another facet of how the data center is
The logic of Intel’s acquisition seems reinvigorating the semiconductor indus-
self-evident now. The data center is the try. Take GPUs, once used for scientific
next growth driver for the chip industry computing and high-end gaming: GPUs
after personal computers and smart- are now being seen as a key enabler of
phones. And it impacts almost every machine-learning functions carried out
facet of semiconductor design. from data centers and cloud servers.
And FPGAs, armed with flexibility
Chips for data centers and reconfigurability, offer even more
Take, for instance, power electronics optimal solutions in the data center
in which analog/mixed-signal design environment. They provide hardware ac-
houses like Maxim Integrated and Vicor celeration for a wide array of applications
are creating low-power systems for data while being paired with CPUs.
centers around initiatives such as the That brings us back to the question
48-V Direct-to-PoL architecture. of why Intel acquired Altera. Intel CPUs
Then there is the emergence of paired with Altera’s FPGAs is akin to gen-
high-speed serial links like PCIe 4.0 and eral-purpose computing coalesced with the
active optical cable (AOC), which have application-specific hardware acceleration
been greatly influenced by the insatiable — a match made in data center heaven?
demand for data capacity in data center Majeed Ahmad
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Publisher’s Perspective 5

Deeply reported news


delivers unique insight
BY W. VICTOR GAO
Publisher and Managing Director,
AspenCore
London, United Kingdom

I
was just in Cambridge, Massachu-
setts, visiting MIT’s ambitious startup
accelerator, The Engine. Answering
MIT President Rafael Reif ’s call to action,
The Engine aims to incubate startups
working on “hard scientific break-
throughs” that deliver transformative
societal impact. As it turns out, much of
that scale of innovation requires hard-
ware design and manufacturing expertise, W. Victor Gao
which most startups lack. And therein
lies one of the most challenging aspects manned moon and Mars explorations.
of commercializing “tough tech” today: A On EE Times, we speak to the movers
whole new class of innovators and entre- and shakers in Grenoble, France, Eu-
preneurs, armed with a profound scientif- rope’s image sensor valley, and examine
ic breakthrough, are constrained by their the feasibility of retooling Europe’s mi-
nescience of electrical engineering. croelectronics industry. Meanwhile, edi-
Yet this challenge is hardly new, nor tors in our Shenzhen, Hong Kong, Taipei,
is it limited to non-electrical engineers. Munich, and Silicon Valley bureaus join
According to AspenCore’s Mind of the hands to present to you an AspenCore
Engineer, the largest, longest-running, Original Series on Made in China 2025:
and most trusted behavioral study of its how China is embracing next-generation
kind, the number-one concern for engi- smart manufacturing technologies and
neers worldwide is that they struggle to what it means for the rest of the world.
keep up with technology change. When In addition to EE Times and EDN,
asked how they coped, they unanimously you can find these articles on publica-
answered that they turned to industry tions such as Electronic Products, Planet
media publications to read and self-train. Analog, EBN, and more. Or, if you prefer
Which brings me to why, at Aspen- to read some of our coverage in their
Core, we believe that good technical original Chinese Mandarin, you can find
journalism is vital to technological them on 21ic.com or follow links on
progress, and rather than degrading English sites to the Chinese edition.
journalism to fit an online strategy of Whether your calling involves finding
ever more clicks and page views, our a better way of probing plasma mem-
first and preeminent concern is to pres- brane to cure a hereditary disease or
ent the best technical data, tools, and inventing a new nanotechnology to run
reporting to you, in the most succinct molecular analysis on a microchip so
language possible, whenever and wher- officials can protect cities from biohaz-
ever you look for insight. ards, at AspenCore, our mission is to
This month, we are excited to present deliver important, deeply reported, and
to you an exclusive, behind-the-scenes deeply mined stories and data about the
look on EDN with NASA and their latest technology and techniques that you need
TDRS-M satellite launch. We interview to do your job. As ever, we look forward
the team at Biosphere 2 about their to your queries, story tips, and feedback.
progress on life sustainment modules for Thank you for your support. ☐
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
ce
fro
an

m
orm

th
erf

is
dp

sin
an

gle
db

co
roa

mp
a-b

on
ltr

en
tu

t
Ge

Ideal for use in Bias Tees, Coilcraft conical inductors offer


flat bandwidth and high impedance to 40 GHz

Coilcraft BCL/BCR Series conical inductors Choose from a rugged, surface mount
operate across a frequency range of 10 MHz package or our flying lead configuration.
to 40 GHz, letting you replace a series And for applications below 6 GHz,
of narrow band inductors with one part. try our high current 4310LC
Both series provide excellent return loss wideband bias choke.
and insertion loss. Their unique conical Learn more and order your
shape optimizes the effects of capacitance, free evaluation samples
maintaining high impedance across your by visiting us online at:
frequency spectrum. coilcraft.com/conicals.

WWW.COILCRAFT.COM
Outlook 7
Innovations impacting products, technology, and applications

High-density, self-healing
silicon could succeed
Li-ion technology
Researchers at the University of Illinois Urbana-
Champaign have developed a technique to
bring the benefits of silicon anodes to batteries
without silicon’s fragility

I
t is no surprise that the next genera- “Silicon is a really interesting
tion of electrically powered portable material,” shared Nancy Sottos, a lead
devices, vehicles, and critical battery researcher on the project and a professor
storage systems will require a signifi- of material science and engineering at
cant upgrade over current lithium-ion UIUC. “It’s also very abundant, relatively
battery technology. Though present-day inexpensive, environmentally sound,
lithium technologies offer high output and silicon can store a lot more lithium
power and good power density, the and so you can get really high theoretical
stress that the batteries undergo during capacities with silicon.”
use causes damage that reduces capacity In current lithium-ion batteries,

Nancy Sottos, professor of materials science and engineering (right), and Scott White,
professor of aerospace engineering at UIUC (left), led a research team investigating methods
for developing self-healing batteries with silicon-based anodes.
IMAGE: L. BRIAN STAUFFER/UNIVERSITY OF ILLINOIS

over time and occasionally leads to fatal the anodes are constructed of graph-
failures that end in fire. ite, which is able to hold fewer elec-
Researchers at the University of Illinois trons than silicon. However, during
Urbana-Champaign (UIUC) are following charge-discharge cycles, graphite only
an avenue of investigation that embraces expands about 6%, whereas silicon
these destructive material tendencies and expands as high as 400%, which leads to
are leveraging silicon-based anodes and cracking and progressive failure. A mar-
self-healing nanoparticle composite bond- ginally successful technique to mitigate
ing agents to bring the benefits of silicon this deterioration is to use segments of
anodes to batteries without the drawbacks silicon in a glue-attached compound
of silicon’s delicacy. anode, though previous gluing attempts
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
8 Outlook
Innovations impacting products, technology, and applications

in research just led to the glue weaken- open and debonds — which would findings that this new technique could
ing and the segments separating. The normally cause a loss of electrical connec- retain 80% of its starting capacity for 400
researchers from UIUC discovered that tion — basically, these bonds are going charging cycles. Compared to lithium-ion
introducing nanoparticle composites to reform and make that interface sound technologies, this is nearly a 400% in-
as bonding agents may not solve the again,” detailed Sottos. “You’ll continue to crease in capacity after 400 cycles.
separation process, but these materials have high performance of the electrode.”
could establish bonds after peeling — The UIUC research team published Researchers are following
ultimately self-healing. the study in the journal Advanced En-
“When it charges, if [the bond] pulls ergy Materials, where they shared their an avenue of investigation
that embraces [lithium’s]
destructive material
tendencies and are
leveraging silicon-based
anodes and self-healing
nanoparticle composite
bonding agents to bring
the benefits of silicon
anodes to batteries
without the drawbacks of
silicon’s delicacy.

Fortunately, or unfortunately, the


UIUC energy storage research team
faces a wide expanse of research and
industrial competition in the energy
storage sector. There are many ar-
eas of investigation for other energy
storage technologies, such as fuel cells
and ultracapacitors, and many groups
researching techniques to enhance the
prolific lithium-ion battery.
“There’s an oversupply of early-stage,
very innovative research in batteries,”
shared Sara Chamberlain, managing
director for Energy Foundry, an approx-
imately $25 million evergreen venture
fund for early-stage energy startups
from Chicago. “And there’s a shortage of
investment capital.”
This is currently leading to many
promising technologies being tabled
due to lack of funding. Nevertheless,
the future does look bright for better
energy storage technologies, as the
competitive landscape for batteries is
likely to yield much higher-capacity,
more reliable, and safer batteries to
power our life on-the-go.
Jean-Jacques DeLisle
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Solid-State Lighting 9

How GaN-on-Si high-power LED


chip technology is changing
luminaire design
Micro-reflectors and
pre-focused LED
manufacturing process
dramatically save space

BY DEBBIE SNIDERMAN
CEO, VI Ventures LLC
www.vivllc.com

T
he latest high-power silicon-based
LED technology from Plessey
Semiconductor is transforming
form, function, and performance when
it comes to luminaire design. Chip-on-
board (COB) LEDs in a typical industrial
luminaire are usually small, but the lens or
reflector assembly and heat sink are much
larger. New GaN-on-Si LED chip technol-
ogy is changing it all, producing luminaire
modules that are only 5.6 mm thick, Using a single monolithic die eliminates the cross-hairs often visible on competing multi-chip
including the optics. A single high-power products.
GaN-on-Si chip can provide 30% more
output through secondary optics than
conventional four-chip solutions.

Silicon substrates are


key to new designs
LEDs at Plessey Semiconductors are man-
ufactured using GaN-on-Si substrates.
Clive Beech, components manager,
explains the many reasons why. “Silicon
is a very good conductor of heat, making
devices easier to heat-sink,” he says. “It’s
economic and available in large quantities
in many areas. Our factory has capa-
bility for CMOS processing on 6- and
8-inch-diameter wafers, a large advantage
over sapphire substrates, which are tradi-
tionally and still run at 4 inches.”
“Silicon is also high-quality and
available with low defect levels, and
6-inch substrates cost around one-tenth
the price of 4-inch sapphire substrates,”
he adds. “This gives us a lower-cost, A packaged Plessey multi-junction GaN-on-Si device.
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
10 Solid-State Lighting

high-performance LED compared to


competitors, allowing us to put LEDs
into new applications such as torches.”
The thermal performance in terms
of thermal resistance is lower on silicon
compared to SiC or sapphire substrates, so
the thermal resistance from the junction
LED surface-mounting point is lower. This
allows Plessey to use higher surface-mount
temperatures at the same junction tem-
perature, which translates into having a
smaller heat sink, saving space. Luminous flux maintenance and color shift over 6,000 hours (LM-80 assessment) measured
In addition to thermal and cost advan- with worst-case conditions of Tsolderpoint = 105°C and IFmax = 1.25 A. The orange line
tages, the surface of the silicon substrate shows very good color stability over the stress period for GaN/Si.
offers another important advantage to
Plessey’s LEDs. Sapphire emits light in all high-power multi-junction LEDs, the same the optics are better tuned to the emitting
directions, so light from LED devices built lumen performance at the same current surface when a single chip is used.”
on that substrate will emit not only from can be implemented with a single chip. They especially help in narrow-beam
the surface but down the body of the LED “This is a huge advantage in lumi- systems like spotlights, portable lighting,
as well, becoming absorbed. Silicon offers naires, especially in spot or other systems flashlights, and 1,000-lumen tactical flash-
surface emission only, so vertical emission that use narrow beams,” says Beech. lights. Having a single, large emitting surface
devices can be created with no absorption. “When a luminaire has a device with provides light that emits over long distances.
It is possible to scale up the area of LEDs four chips in a high-power package, they The new high-power LUCIAN
without losses in the bulk and even see are traditionally put in a 2 x 2 array with PLW7070 single-chip LEDs can have
performance improvements. no chip at the center. Putting this light larger chip dimensions compared to oth-
Another feature of using a silicon sub- through narrow-beam optics projects a ers produced on sapphire substrates and
strate is that multi-junction LED devices crosshair in the middle onto a surface are used for portable torches and lamps,
can be integrated with its circuitry on a from the four chips. With a single chip spot-lighting, down-lighting, high-bay,
single chip. This is a distinctive advan- and a single light-emitting surface, there and floodlighting.
tage for Plessey, who is the only one is no crosshair pattern.” The operating life of Plessey’s GaN-
currently offering LEDs integrated with The single monolithic high-voltage on-Si LEDs have much better lumen
optics in the industry. device produces a much better optical maintenance. Competitor LEDs’ light
far-field image. It is also centered, being output degrades 7% to 10% over 6,000–
Changing luminaire design that the center of the emitting surface is 7,000 hours, while theirs only reduces by
Competitors use an array of smaller LEDs at the center of the package. So there is 3% over the same time.
in their high-voltage packaged devices. For also superior central beam candela power
example, 12-V LEDs use four 3-V LEDs compared to a multi-chip solution. The New process for creating
in series. With Plessey Semiconductor’s peak intensity at the surface is greater pre-focused LEDs
than those that are not optically Plessey Semiconductors’ patented
centered. “One customer uses Stellar LED manufacturing process was
our chip to produce a system developed in 2015. It involves patterning
with a beam angle of only two GaN devices on a silicon substrate and
degrees,” says Beech. “Overall, mounting a second silicon substrate
on top, then flipping the
entire structure over and
removing the back side to
reveal a new substrate. This
allows holes to be etched in
the substrate that reach the
LED below, effectively cre-
ating LEDs that are layered
and sandwiched between
two substrates.
A smaller heat sink enables further size reductions in The most important
the end fixture. Stellar ORION product. feature of this process is
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Solid-State Lighting 11

array of micro-reflectors around each of as down lights, small heat sinks are im-
the LEDs. The spotlight’s total thickness, portant. High-power LEDs are similar to
including the integrated reflectors, is 7 other power devices: Power goes in, light
to 8 mm thick, the same as the diameter and heat come out. For a given level of
of the reflector and much smaller than heat generated, the thermal path is key to
other technologies. what the junction temperature will be.
Some spotlight applications require a As a starting point for designing the
cover plate for environmental protec- heat sink, it must not exceed a 135°C
tion, but there is no lens. The optics are junction temperature. At the maxi-
The PLW7070 multi-junction chip (unlit). designed and built into the unit itself.
Integrating the optics into the LED
the ability to create micro-mirrors that system itself is unique in the industry,
pre-focus the light coming out of the and currently, Plessey Semiconductors
LED at the chip level as part of the LED is the only one offering this advantage.
processing. The process doesn’t remove Samples of its Stellar Orion product are
certain areas from one of the substrates currently shipping with a range of 25-
and fills the holes to form the micro-mir- and 60-degree beam angles.
ror. The end result is that the LED pro- “These are now our standard prod-
duces inherently narrow beams that don’t uct and are more attractive and look less
require secondary optics. The pre-fo- industrial than the typical spotlights they On-chip optics.
cused narrow-beam LED system saves are replacing,” says Beech. It is a stepping
money, size, and the need for additional stone toward a fully integrated solution in mum operating current, given the low
optics and can become much smaller and which the LEDs have built-in collimated thermal resistance of the substrate, the
less expensive in time. beams and will be much smaller, anticipat- solder-point temperature at which the
ed to be available in the next year or so. At device is mounted to the circuit board
New reflector designs: COB that point, the micro-reflectors wouldn’t will be about 20 degrees below that, or
arrays with micro-reflectors be needed, and the total thickness of the about 115°C.
Plessey’s Stellar Orion LED beam-form- devices integrated into the LED chip itself “We face our heat sinks to the outside
ing module is a COB version of these would be on the order of a few hundred air,” says Beech. “The thermal resistance
integrated optics for directional lighting microns. “The current standard product is of the heat sink to the ambient environ-
applications that Beech says has won sev- like an analog version of what’s coming at a ment is a strong function of the specific
eral awards. Normally, a 15- or 16-degree much larger scale,” he says. design and surface area. Because we run
spotlight would have a large reflector and at higher solder-point temperatures, we
a mirror on the front surface, but theirs Evolving heat sinks can use smaller heat sinks and still have
doesn’t need them. Instead, it uses an In devices with space constraints, such good performance.” ☐

0201HL Wirewound Chip Inductors


The Highest Inductance in an 0201 Package
• Inductance values from 22 to 51 nH
• Optimized for 700 MHz LTE and 5G applications
• Twice the Q factors of thin-film multilayer
technology ... up to 62 at 2.4 GHz
Full Specs & Free Samples @ coilcraft.com
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
12 Memory ICs

How ever-evolving data centers


will reshape SerDes and
memory designs for the better
High-speed serial links, memory, and acceleration are playing a critical
role in meeting the never-ending demand for greater data capacity
become a reality when single-lane 112- for SerDes designers. In addition to
BY MOHIT GUPTA
Senior Director, Product Marketing; Gbps links hit the market. maintaining signal integrity, engineers
AMIT KEDIA, Senior Field Applications This technology will allow larger lane must contend with a stringent set of
Engineer, Rambus Inc. bundles, providing 1-TbE or 1.6-TbE design requirements, such as architecting
www.rambus.com links with 10 or 16 lanes, respectively. next-generation silicon within the same

N
Data centers maintained by Amazon, power envelope without the support of
etwork infrastructures and their Facebook, and Google are moving to Dennard scaling.
supporting data centers are rapid- 100-GbE connections in 2017 and are Nevertheless, the industry is eyeing
ly evolving to meet the challenges expected to begin purchasing 400-GbE accelerated innovation to facilitate the
of our increasingly connected world. systems by 2019. transition from 28-Gbps to 56-Gbps to
Indeed, petabytes of data are now generat- As such, it is critical for the current 100-Gbps single-lane.
ed by a wide range of devices, systems, and generation of 56-Gbps SerDes PHYs to Regarding the wireless market, mobile
IoT endpoints such as vehicles, wearables, meet the long-reach backplane require- operators and vendors have already
smartphones, and even home appliances. ments for the industry transition to 400- embarked on R&D initiatives to develop
This article will explore how high-speed GbE applications. It allows the SerDes 5G, which will be capable of supporting
serial links, memory, and acceleration are PHY designs to scale to speeds as fast next-generation self-driving cars, UHD
all shifting into high gear to accommodate as 112 Gbps, which are required in the video, cloud robotics, 3D holographic
the demand for ever-faster speeds. networking and enterprise segments. telepresence, and remote surgery with
Take the case of enterprise server haptic feedback.
Serial links: Ethernet and 5G racks that are moving from 100 GbE to The very first standardized deploy-
The 400-Gigabit Ethernet (400 GbE) 400 GbE and beyond. Ethernet is moving ments of the technology are expected
and 200-Gigabit Ethernet (200 GbE) faster than ever, and there is already a to be commercialized by 2019 with the
are slated for official release by the IEEE forum for 112-Gbps SerDes, which is go- 3GPP’s initial 5G specifications slated to
P802.3cd Task Force in December 2017. ing to be critical in driving the 800-GbE be implementation-ready by March 2018.
Although the official IEEE roadmap has standard forward. According to a new report by Research
yet to precisely detail what lies beyond Supporting faster Ethernet speeds and Markets, the 5G network infrastruc-
400 GbE, doubling to 800 GbE will likely presents a distinct set of challenges ture market is projected to “aggressively
grow” at a CAGR of nearly 70%, eventu-
ally accounting for $28 billion in annual
spending by the end of 2025. These
infrastructure investments will be accom-
panied by annual shipments of up to 520
million 5G-capable devices.
Perhaps not surprisingly, the impend-
ing shift to 5G has prompted wireless
base station equipment manufacturers to
design new architectures and systems to
meet the design requirements by pushing
high-speed interfaces like CPRI (to con-
nect RF components) from 24G to 48G
and JESD204 (to connect data convert-
ers) from <15G to 32G.
Exponential data growth trend from IDC study. With interconnect speeds fast
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Memory ICs 13

with multiple entities, including the


foundry and packaging house. So a care-
ful design of the entire system — includ-
ing SoC, interposer, DRAM, and package
— is essential to ensure high yield and
proper system operation.
In addition, having a high-yielding
module is a critical element of keeping
costs in check, given the number of expen-
sive components, including the SoC, mul-
A view of network peripherals. tiple HBM die stacks, and the interposer.
However, even with these challenges, the
approaching 56G and beyond, SerDes ing (HPC), server, networking, and client advantages of having increased memory
PHYs will play an important part in the applications. As system designers look bandwidth and density closer to the CPU
development of wireline and wireless to move higher bandwidth closer to the clearly improves overall system efficiency
networking infrastructure. Therefore, CPU, HBM2 offers an opportunity to sig- for server and networking systems.
system architects should carefully nificantly expand memory capacity and
consider the pros and cons of design maximize local DRAM storage for wider Acceleration: FPGAs and GPUs
tradeoffs around die area and power. throughput in data center environments. Along with high-speed serial links and
Deciding on a specific set of tradeoffs The HBM DRAM architecture memory, GPUs and FPGAs have also played
for high-speed serial links may very well increases system memory bandwidth by a significant role in helping to shape the
be one of the most complicated parts of providing a wide interface of 1,024 bits modern data center. For instance, GPUs
the entire design process. It’s difficult to the SoC design. More specifically, sec- offer optimal performance, power efficiency,
to maximize operational efficiency and ond-gen HBM offers a maximum speed and compute density for many SIMD-style
performance simultaneously. of 2 Gbps or a total bandwidth of 256 workloads with the use of a large number of
Gbytes/s. Although the bit rate is similar processing pipelines that also support some
Memory: DDR5 and HBM to DDR3 at 2.1 Gbps, the eight 128-bit amount of programmability.
Memory architecture is also evolving to channels provide HBM with approxi- First deployed in supercomputers and
meet the demand for increased band- mately 15 times more bandwidth. data centers for high-performance and
width in both the wired and wireless The HBM modules are connected to the scientific computing applications, GPUs
markets. For example, JEDEC recently SoC via a silicon or organic interposer. A have since achieved wider adoption as
confirmed that its upcoming DDR5 short and controlled channel between the machine-learning applications have
memory standard will offer improved memory and the SoC requires less drive grown in popularity. Coupling a large
performance with greater power efficien- from the memory interface, thus reduc- number of parallel processing pipelines
cy as compared to previous-generation ing the power when compared to DIMM with the highest bandwidth memories,
DRAM technologies. The DDR5 speci- interfaces. Moreover, because the interface GPUs dramatically improve the time to
fication will double the bandwidth and is wide, system designers can achieve very train neural networks as they process
density over DDR4 along with delivering high bandwidth with a slower frequency. training data at unmatched speeds.
improved channel efficiency. As expected, the implementation of For other types of workloads, FPGAs
Meanwhile, high bandwidth memory 2.5D technology in HBM2 systems adds offer a more optimal solution than GPUs
(HBM)-based DRAM has been adopted numerous manufacturing complexities, due to the broader flexibility and recon-
for graphics, high-performance comput- requiring PHY vendors to work closely figurability they enable to meet the vary-
ing demands of different applications.
And while traditional CPUs offer the
ability to run complete workloads of any
type, FPGAs provide application-specific
hardware acceleration and offload capa-
bilities that can be updated over time.
Put simply, the flexibility afforded by
FPGAs makes them a particularly powerful
solution when paired with CPUs, allowing
the benefits of general-purpose computa-
tion to be combined with those of hard-
ware-accelerated processing for time-con-
High-speed SerDes in 5G infrastructure. suming and power-hungry tasks. ☐
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
14 Cover Story

Smart vehicles launch the


autonomous moonshot
Full autonomy may be years away, but the and ranging, to name a few (Fig. 1). Geo-
partnerships and technological building blocks to graphically, Asia-Pacific will dominate in
semi-autonomous, followed by Europe
help designers get there are developing rapidly and the U.S., according to the report.
A report by McKinsey points to a
BY PATRICK MANNION units in 2016 to 7.84 million by 2021. The move away from direct car ownership
Contributing Editor number of autonomous vehicles will rise to car sharing, or using the best mode of

A
from 0.18 million in 2025, reaching 1.01 transport for the task at hand, on demand:
pple CEO, Tim Cook, said it best: million by 2030. For the purposes of the “In the U.S., for example, the share of
Autonomous vehicles are the report, the firm segments semi-autono- young people (16 to 24 years) that hold
“mother of all AI projects.”1 Despite mous as having adaptive cruise control, a driver’s license dropped from 76% in
the technical challenges, or maybe because 2000 to 71% in 2013, while the number of
of them, semi-autonomous and autono- car-sharing members in North America
mous vehicles are on the march across and Germany has grown by more
Asia, Europe, and then the U.S. than 30% annually over the last
Tell that to engineers im- five years.”
mersed in current automotive The combination of
or mission-critical hardware/ changing user patterns, the
software development, and preference for using mobile
they scoff. Tell it to sci-fi buffs devices versus driving, and
and enthusiasts, and they say the growth of autonomous
it’s just a matter of when and how. features has sparked a flurry of
Tell it to a friend or family member, activity in the smart-vehicle arena.
and they say, “No thanks.” Google’s Waymo, the current leader in
The skepticism of engineers and con- self-driving technology, has partnered
Fig. 1: Vehicles will get smart as soon as
sumers is well-founded, given the matrix with Lyft while it does battle with Lyft’s
designers and developers get even smarter
of problems associated with implement- direct competitor, Uber, which it accuses
about how to integrate all sensing options.
ing semi-automated vehicles, from re- of illegally acquiring technological
al-time object detection and recognition secrets. At the same time, Lyft has a
to communication and security. Add in Autonomous vehicles partnership with GM to start deploying
news reports of failures, and it’s hard to are the “mother of all AI thousands of self-driving electric cars in
convince users of their benefits. 2018. Along the same vein, Waymo has
However, the sheer magnitude of projects.” Despite the partnered with Avis to have Avis take
the challenge is also its attraction, as is technical challenges, or care of its fleet of autonomous vehicles.
the possibility of saving lives once the maybe because of them, Apple has done likewise with Hertz as it
problems are worked out. There’s also the plans its own autonomous vehicle moves.
attraction of the monetary rewards to be semi-autonomous and These moves are in parallel with those
made by solving the matrix of technical autonomous vehicles are of almost every automotive company
challenges, as well as opening up new on the march across Asia, — including BMW, Audi, Chrysler, Mer-
IMAGE: (FIG. 1) UNIVERSITY OF TEXAS, AUSTIN

ways of thinking of automobiles — as cedes-Benz, and Ford — that is rapidly


sources of data and a conveyor of new Europe, and the U.S. expanding the role of technology, to the
services — instead of simply seeing them point that the way forward for growth for
as a mode of transportation. lane keep assist, intelligent park assist, the automobile industry may be the addi-
adaptive front lighting, blind-spot detec- tion of services on top of the transporta-
Opportunity spurring tion, and emergency brake assist, among tion function.
accelerating activity other features. Under autonomous, it Ford has already pointed in this direction
According to research firm Market- includes radar, LiDAR, ultrasonic, cam- with the recent hiring of Jim Hackett as its
sandMarkets, semi-autonomous vehicle eras, engine control units, fusion sensors, CEO. Hackett was leading the Ford Smart
shipments will increase from 3.17 million steering assist, control processing units, Mobility LLC group, exploring ways to add
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Cover Story 15

Fig. 2: The InnovizOne is a solid-state LiDAR product for vehicles that can detect at distances up to 200 m, with a depth accuracy of <10 cm.

value to the “smart mobility experience.” ments since 2005. LiDAR’s advantages While Velodyne initially seemed to
For companies like Ford, the oppor- over radar include the ability to grab a own the LiDAR space, the introduc-
tunity lies not just in selling cars, but high-resolution 3D map of the surround- tion of smaller, low-cost solid-state
also in acquiring data about, and selling ing area, but it has to deal with ambient and micro-electromechanical systems
services to, the vehicle occupants. The light variations, including those caused (MEMS) LiDAR in the $250 range in
vehicle itself becomes simply a conduit by fog. Radar has the advantage of being late 2016 has changed the calculus. A
for those services. This echoes recent RF-based, so it can see objects, regardless good example of the new device is the
statements by Intel’s Brian Krzanich, who of ambient lighting, and it can also see S3 from Quanergy. Now a full 360° view
flat out said that it’s all about the data and “around” objects using radio reflections. around a vehicle can be generated using
its analysis. “Data is the new oil,” he said.
The types of data vary: There’s data
that’s worth mining on the user and their
driving habits, location, and physical and
mental states. There’s also data from the
vehicle to other vehicles (V2V) and to
infrastructure (V2I) as part of the emerg-
ing applications for advanced driver
assistance systems.
How that data is to be secured and
used is an issue for security and privacy
experts. How that data is acquired and
IMAGE: (FIG. 2) INNOVIZ; (FIG. 3) LEDDARTECH

communicated is where designers of au-


tomotive systems come in, along with the
components that go into those systems.

Object detection and


ranging options
For autonomous vehicle object detection
and ranging, it seemed that LiDAR was
the way forward, with Velodyne leading Fig. 3: LeddarTech packages its proprietary LiDAR ranging and detection technology into an
with its familiar “coffee-can”-style instru- IC, but it also offers subsystems, such as the Vu8 eight-sensor segment module.
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
16 Cover Story

four $250 solid-state LiDAR devices,


one placed at each corner of the vehicle,
for a total cost of $1,000. This compares
favorably to the $70,000 to $85,000 for
a classic Velodyne system, with its awk-
ward mechanical rotating can.
However, Quanergy is not alone, with
many startups offering other relatively
low-cost LiDAR options. Innoviz, out
of Israel, has promised $100 solid-state
LiDAR by the end of 2018.
To compare solutions, designers need
to look at field of view (FoV) in the ver-
tical and horizontal direction, range for
a given reflectivity, and resolution. The
latter two are inversely related: The lon-
ger the range, the harder it is to achieve a
given resolution.
For Quanergy’s S3, the FoV is 120°
both horizontally and vertically. It has a
range of 150 m at 8% reflectivity, and at
100 m, its accuracy is ±5 cm. The sensor
measures 9 x 6 x 6 cm. Innoviz’s initial Fig. 4: The VERA-P173 from u-blox provides V2X communications for smart vehicles using
product, the InnovizOne, measures 5 x IEEE 802.11p, a variant of Wi-Fi for WAVE.
5 x 5 cm, has a range of 200 m, an FoV
of 100° horizontal and 25° vertical, and teristics now achievable as it moves into along with safety, include better traffic
a depth accuracy of 2 cm (Fig. 2). the millimeter-wave bands in the 7x-GHz flow, leading to lower fuel consumption.
While Velodyne initially eschewed range (see Page 22, “Sensors and trans- The underlying communications
solid-state technology, it has since come up ducers go wireless and get smarter”). This interface for V2X is IEEE 802.11p, which
with its own version of a solid-state LiDAR. makes for smaller devices, and the wider is a variant of Wi-Fi for wireless access
The Velarray sensor measures 125 x 50 x 55 bandwidth allows for finer resolution. in vehicular environments (WAVE) in
mm and can be embedded into the front, The improvements in both LiDAR the 5.9-GHz band. It is optimized for
corners, and sides of vehicles for auton- and radar do not mean that one will fast response times (low latency) in the
omous or ADAS-enabled vehicles. The win at the expense of the other. Instead, tens-of-milliseconds range, as well as
Velarray uses Velodyne’s own ASIC design the goal now is to use both, along with longer range.
and features a horizontal and vertical FoV cameras and other available sensing To design for V2X, the options are
resolution of 120° and 35°, respectively, technologies, to have a more compre- to either design from scratch or to opt
with a range of 200 m. It quotes price as hensive and optimized solution for for a module. The latter saves time and
being “in the hundreds of dollars.” maximum safety and performance at cost with respect to design and meeting
There are other LiDAR options for optimum cost. regulatory compliance. A good option
designers, too. For example, LeddarTech That said, the optimum smart-vehicle is the VERA-P1 series from u-blox. It
has broken down the ranging problem design doesn’t operate in a vacuum. measures 24.8 x 29.6 x 4.0 mm and is
and focused on the algorithms, which it compliant with WAVE and ETSI ITS G5
has implemented in a proprietary Led- Communication makes vehicles for the U.S. and Europe. It has a range
darCore IC, though it also offers a full smarter of more than 1 km (line of sight), with
system (Fig. 3). Though the on-board sensing and safety a receive sensitivity of –97 dBm and
Designers can take the IC and develop mechanisms in a smart vehicle are ex- an output power of –10 to 23 dBm. It
their own custom-value add for specific tensive, all smart vehicles must be able operates over the temperature range of
applications by adding a light source, to communicate with and incorporate –40°C to 95°C. ☐
optics, and detector. input from their surroundings. This in-
IMAGE: (FIG. 4) U-BLOX

While LiDAR is dropping quickly in cludes V2V, V2I — such as traffic lights, Reference:
cost, making semi-autonomous and auton- emergency services, and location-based 1: Bloomberg Technology, “Tim Cook
omous vehicles more cost-effective, radar services — or both (V2X). The challenges Says Apple Focused on Autonomous
is also making headway, particularly with here are many, but chief among them is Systems in Cars Push”: https://bloom.
respect to the size and performance charac- latency. The advantages of getting it right, bg/2rVRzed
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Special 17

Understanding and solving


industrial robot design
challenges
Designing advanced next-generation robots for industrial applications,
especially where they collaborate closely with humans, poses safety and
performance challenges that can be met with proper attention to sensors,
power, and communications
BY TOBIAS PUETZ The expanding role of robots Example sensors to help avoid collisions in-
Systems Engineer, Factory Automation For over 50 years, robots have brought clude ultrasonic and LiDAR. The controller
and Control, Texas Instruments productivity, cost-efficiency, and im- is located inside with a wireless interface to
www.ti.com proved safety to repetitive task perfor- a central control center.

F
mance. Though their roles have expand-
or more than half a century, robots ed greatly, they can be generally placed Robots can be generally placed
have played an ever-increasing role into three main categories: industrial,
in manufacturing, successfully trans- logistical, and collaborative (Fig. 1). into three main categories:
forming industries ranging from automo- Each has their respective sensing and industrial, logistical, and
biles to electronics to consumer goods. As safety requirements and best practices.
collaborative. Each has
their adoption continues, they are evolving Industrial robots are fixed in place to
to become more connected and intelligent handle tasks such as welding, painting, their respective sensing and
while working more closely and collabora- picking and placing, and assembly. The safety requirements and best
tively with their human counterparts. controller is usually either in the base of
practices.
This presents a number of challenges the robot or next to it. They are designed
for designers who are being asked to add to be quick and accurate with no direct
greater functionality, flexibility, range of human contact. For safety, they are Collaborative robots work next to hu-
motion, speed, and precision while, at the usually isolated inside fences with added mans, and so they are the most complex.
same time, improving safety and reliability. protection, such as floor mats that trip They need particular attention to safety
This feature will describe three differ- the power when stepped on. and need to perform fine-tuning tasks or
ent classes of robots: industrial, logistical, Logistical robots are mobile, are hold an object while a worker inspects it.
and collaborative. It will then focus on typically used to fetch goods or transport They require an array of sensors for col-
the design challenges, what parameters loads in warehouses, and may operate with lision avoidance and failsafe mechanisms
are most important to the system, and humans present. As such, they usually to turn off the robot should anything go
how to go about selecting components to have a handler and move within a certain wrong. They are typically fixed in place
help meet those challenges. environment using sensors and mapping. but can also be mounted on a vehicle.

Fig. 1: Robots for industrial and factory automation can be classified into three main categories: industrial, logistics, and collaborative.
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
18 Special

It’s clear that for robots to operate in


these increasingly complex ways, they
must be able to process a great deal of
sensing data about the environment,
communicate with each other and with
centralized control units, and perform
control functions that adapt to environ-
mental changes and keep them from
harming humans.

Technology requirements for


industrial robotics
The three areas of technology that play a
major role in robotics are: communica-
tion, motor control, and sensing. Robot
systems are often built together out of Fig. 2: Features such as the PRU-ICSS exist as a hardware block inside a processor to support
products from several different manu- multiple industrial communications protocols.
facturers. Therefore, it is necessary that
all of these products can be connected cation controllers a requirement. processor with single or multiple cores to
either to each other or to the main robot Processors for such control need to ensure that the right balance of integra-
controller. For a design engineer, this provide flexible peripherals, connectivity, tion, connectivity, and performance is
means that a system must support multi- and unified software support to cover available for every application.
ple protocols, making flexible communi- a wide set of applications. Look for a Flexibility is often achieved using

Fig. 3: A good 3D time-of-flight (ToF) image sensor combines ToF sensing with an optimized analog-to-digital converter and timing generator
with a QVGA resolution to detect people on manufacturing lines and distinguish between objects and between objects and humans.
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Special 19

external FPGAs or application-specific industrial, logistical, or collaborative, time-of-flight (ToF) system that can, for
ASICs, but there are other options to sensing is playing an increasing role. example, be mounted onto the wall or
consider, such as a single-chip solu- For collaborative applications in which ceiling and be used for people recogni-
tion that combines a processor with a sensing systems must be able to not only tion at manufacturing lines (Fig. 3). The
programmable real-time unit inside the detect if a human is around but also pre- system returns a point cloud that can be
industrial communication subsystem vent any possible collision, sophisticated further processed in software.
(PRU-ICSS) (Fig. 2). sensor technologies are a core require- This gives the end user the ability to
These exist as a hardware block inside ment. They must be able to cover a wide not only detect objects but also distin-
the processor and support multiple range of up to several meters while offer- guish between steady objects, robots,
industrial protocols like EtherCAT and ing resolutions down to centimeters or and humans. Furthermore, a wide field
ProfiNET, among others, which enables even millimeters. Furthermore, multiple of view of 87° diagonal, 74° horizon-
designers to build in flexibility and inter- sensor systems are needed to cover the tal, and 59° vertical can be covered
face to every part of the robot system. complete range from near- to far-field. at an operating range of 4 m. Lastly,
Meeting these design requirements is such a system can also be used in dark
Efficient motor designs needed difficult. Common challenges are vari- environments because of the usage of
As more motors are used in a robot able sensing ranges and resolutions, dirty near-infrared (NIR) lasers emitting at a
system, motor efficiency becomes critical wavelength of 855 nm.
— not just for large load-carrying motors,
As more motors are used
but also for smaller, collaborative robots Building better robots for the
as they increase in usage. Collaborative in a robot system, motor integrated factory
robots often have the motor control efficiency becomes critical — As manufacturing continues to become
boards integrated into the robot arm; more highly integrated at all levels, ro-
not just for large load-carrying
that is why designers need to pay special bots will play an ever-increasing role in
attention to the thermal management. motors, but also for smaller, carrying out a wide variety of assembly
Therefore, it is not only necessary to select collaborative robots as they tasks that increase production and make
components that are small in size but also the workplace safer for human beings.
increase in usage.
to use switches with low switching losses. Traditional industrial robots, logis-
Gallium-nitride (GaN) technology is tical robots, and collaborative robots
an important design asset here because it environments, and the difficulty of dif- have their jobs to do. So, too, do robot
greatly reduces switching losses, thereby ferentiating between humans and other developers as they look for solutions
enabling faster switching speeds while robots working in the same environment. that enable accurate, safe, cost-effective
simultaneously facilitating thermal The solution is to use multiple sensors operation from all of them. To that end,
management. To simplify and speed the systems: capacitive sensing for the near- designers should also ensure:
design process, designers should look for range (up to 15 cm), infrared sensing
modules that provide a complete solution for middle-range (15–150 cm) and laser • A high-efficiency, high-voltage power
with optimized layout and efficiency, sensing for long-range (>150 cm). How- supply with circuit protection and low-
along with minimum electromagnetic ever, laser systems come along with a lot noise emissions
emission and noise for compliance with of safety restrictions. • Characterization for an extended tem-
industrial standards. These challenges can also be solved perature range
Also, support should include EVMs, by using new and advanced sensing • Fast, precise analog-to-digital and digi-
development boards, and a quick-start solutions, like millimeter-wave radar in tal-to-analog signal conversion
toolset to speed design. For specific the 76- to 81-GHz bands. Sensors operat- • Reinforced isolation to meet industrial
functions, such as accurate control of ing in this region can detect objects at safety standards
precision drives, a vendor’s reference a range of up to 84 m, with a resolution • Control redundancy for safety-criti-
design library may already include an of 37 cm. Depending on the system cal applications when combined with
appropriate design. requirements, these parameters can also other ICs
Using well-supported reference de- be tweaked for shorter range and higher
signs speed time-to-market by reducing resolution. In many cases, it’s smarter to work
the chance of design errors while also Millimeter-wave ranging sensors with a vendor with a wide degree of
freeing up designers to focus on their also allow velocity measurement of the experience in the design and application
design’s core differentiation and value. detected object, and because it is RF, it of robotic technology in industrial appli-
is robust in the face of environmental cations. It’s highly possible that they’ve
Better sensing in challenging conditions such as dust, smoke, fog, rain, either solved for the various challenges
environments sunlight, or darkness. of a design or have the expertise to help
Regardless of whether robot designs are Another ranging option is a 3D uncover a new solution. ☐
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
20 Product Roundup
Packaging, Cabinets & Enclosures

Enclosure for pluggable connector front-pluggable cards and the optional rear transition module
(RTM) section. Next, it features standard 600- or 1,200-W mod-
systems boasts UL 508A ular power supplies for a wide range of voltages and amperage
certification levels commonly used in OpenVPX and other backplanes.
Pixus also offers power interface boards as per PICMG 2.11
Certifications are of and VITA 62 specifications to accept various pluggable power
vital significance in supplies. Here, it’s worth mentioning that Pixus offers 3U and
enclosures for industrial 6U backplanes in OpenVPX, CompactPCI Serial, VME64x,
control equipment be- and legacy CompactPCI and VME formats.
cause they maintain the Pixus Technologies: www.pixustechnologies.com
original environmental
protection rating even
after the cutout. Take,
for instance, Phoenix
Steel
Contact’s Heavycon con- enclosure is
nector enclosure series,
which conforms to the
available in
UL 508A standard. 18 sizes
The Heavycon connector enclosures are also certified for Polycase has launched a
UL Type 4, 4X, and 12 protection categories for panel feed- new line of NEMA-rat-
through connectivity in electrical enclosures. ed steel enclosures
The UL 508A certification is critical for North Amer- that protect industrial
ican cabinet builders who produce UL-certified electrical equipment against circu-
enclosures for industrial control applications. That’s because lating dust, falling dirt, rain,
enclosures complying with UL 508A standards must use and hose-directed water. The SB Series steel enclosures comply
components that have been evaluated and listed by the UL. with the UL 508A standard and meet NEMA 1, 2, 4, 4X, and 12
Phoenix Contact: www.phoenixcontact.com requirements along with IP65 and IP66 ratings.
The new steel enclosures are available in 18 sizes ranging
from 8 x 6 x 6-in. up to 24 x 24 x 10-in. dimensions. And
Test enclosure aided by Polycase can customize the size of SB Series enclosures using
removable sidewalls CNC cutouts.
Every enclosure unit comes with an internal panel, external
Pixus Technologies has introduced a development enclosure mounting brackets, latch-key, and grounding wire. A smooth
for 3U, 6U, or custom-sized boards with sidewalls that can be gray powder coat covers the 16-gauge steel housing.
removed. That allows a system to be enclosed for thermal test- Then there are features such as a hinged door with keyed
ing. And the walls can be taken out for easy access to boards quarter-turn latch, watertight gasket, and bonding studs for
inside the chassis. grounding in both the base and the door. The SB Series steel
The VPXD1000R enclosure holds up to eight backplane enclosures can be purchased on the Polycase website.
slots at a 1.0-in. pitch. And card guides can be adjusted in Polycase: www.polycase.com
0.2-in. increments to accept various slot pitches. Further-
more, Pixus offers
removable con-
duction-cooled
Plastic enclosure targeted at low-
card guides, profile applications
which allows Polycase has added a new size to its DC Series of heavy-duty
IEEE 1101.2 plastic enclosures for electronic systems. The DC-57 enclosure
conduction-cooled — measuring 8.25 x 5.0 x 1.5 in. — has been developed for
modules to be applications that require a low profile.
tested next to other It’s molded from
air-cooled cards black ABS plas-
inside the same tic with a UL94-
enclosure. 5VA flame
The development rating. And
enclosure has remov- the DC-57
able sidewalls for both enclosures

AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS


Product Roundup 21
Packaging, Cabinets & Enclosures

are available with or without being molded on surface-mount- and NEMA 6 ratings using
ing flanges. Other features include bass inserts, PCB mounting a latch or screw-locking
bosses, choice of textured or smooth cover, an optional internal method. And the polycar-
panel accessory, and an optional gasket. bonate (PC) enclosure can
Polycase also offers customization services, including withstand UV exposure,
custom CNC machining and digital printing on the DC-57 rapid temperature swings,
enclosures. These new plastic enclosures can be purchased from and extreme temperatures
the Polycase website. ranging from –40°C to
Polycase: www.polycase.com 85°C.
Phoenix Contact’s
new case also offers a
Test socket boosts contact PCB lock, which prevents
falls and ensures easy
protection with plunger matrix maintenance. And it can
Ironwood Electronics has unveiled a QFN socket design for the be readily mounted on a
9.1 x 14-mm package size. The SMP-QFN-8019 socket employs pole or a wall without compromising its ease of PCB repair or
high-performance elastomer that facilitates very low induc- replacement.
tance, high endurance, and wide temperature range. The electronic enclosure for rugged applications has inte-
grated connectivity options with one or two M12 inserts, as
well as an option for a cable gland. Alternatively, a blank face-
plate can be modified to accommodate any sealed connector
that fits within the housing profile.
The integrated connection technology outlined above simpli-
fies packaging devices intended for control on vehicles, rugged
industrial wireless applications, and even small-cell installations.
Phoenix Contact: www.phoenixcontact.com

The test socket operates at bandwidths of up to 44 GHz with F R E E C ATA L O G & B R O C H U R E

STAMPINGS
less than 1 dB of insertion loss. Next, the contact resistance is
typically 15 mΩ per pin, and the socket connects all pins with
44-GHz bandwidth on all connections.

& WASHERS
The SMP-QFN-8019 socket, mounted on the target PCB
without soldering, uses mechanical hardware for the mounting
operation. The socket is constructed using Torlon material with
a cutout to accommodate RF traces on the top side of the PCB.
The socket comes with a protective plunger matrix — a
gold-plated copper cylinder — which sits on top of the conductive
silver columns. The plunger matrix protects the conductive silver
columns from contamination due to various solder ball interfaces.
A quickly replaceable plunger matrix also enables minimal
downtime during final production test. The silver column with
plunger matrix contact is rated for more than 500K cycles.
Ironwood Electronics: www.ironwoodelectronics.com

bokers.com
Environmental case features Call (888)-WASHERS 612-729-9365 [email protected]

integrated connectivity options


The new environmental case system (ECS) from Phoenix Con-
tact protects electronics from potentially damaging elements
commonly found in outdoor applications and extreme indoor
environments. 3D-PRINTED PROTOTYPES NOW AVAILABLE
The two-piece, sealed environmental case facilitates IP66, IP67,
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
22 Product Trends
Insight into current product developments

Sensors and transducers go


wireless and get smarter
Sensors are where the IoT and safer vehicles begin, and wireless
connectivity, combined with higher levels of functionality, is helping to
deploy sensors everywhere

BY PATRICK MANNION easements, low power is just as critical,


Contributing Editor with a 10-year battery lifespan being the

R
common reference point.
ecent developments in sensors and LoRaWAN is supported by the LoRa
transducers show increasing em- Fig. 1: The BMA456 and BMA243 low-power Alliance, which recently crossed the
phasis on connectivity for the IoT, MEMS acceleration sensors from Bosch 500-member point. It operates in unli-
intelligence for more integrated, stand- Sensortec are specifically designed for censed bands versus Sigfox’s licensed-band
alone solutions, and higher performance motion and fitness tracking. operation. It has been held back by limited
in terms of precision and accuracy. silicon availability, with Semtech being
The applications driving these more standalone processing and control the only provider of the transceiver and
developments include asset and device features (Fig. 1). Microchip supplying a module based on
tracking and sensing, wearables, auto- The BMA456 and BMA243 include that transceiver. However, STMicroelec-
motive, and smartphone applications an optimized step counter for wearables tronics has joined the group and is already
such as virtual reality and augmented integrated directly into the sensor, with producing silicon. Renesas also joined, and
reality. In particular, radar-based ranging no additional external microcontroller it is expected to do likewise.
and position sensing for automotive has required. This helps designers’ cost and Even as LoRaWAN ramps up the
made marked improvements, thanks to power consumption within acceptable number of chip suppliers, the two inter-
a technology shift to the 76- to 81-GHz limits, reducing design complexity and faces will continue to offer designers a
millimeter-wave bands. time-to-market. good connectivity option, as does Nar-
The integrated packaging of sensors Key parameters include 16-bit resolu- rowband-IoT from cellular operators.
is also a key enabler for developers who tion, a noise figure of 120 µg√Hz, and total In the meantime, some interesting
need to respond quickly to customers’ offset over its lifetime of only 20 mg. The sensor options have come about, such as
needs, particularly in consumer applica- power consumption for the step-counter the LoRaWAN sensor transceiver from
tions. Bosch Sensortec GmbH is focused function is under 30 µA and the package Adeunis RF (Fig. 2).
on the consumer arena, and according measures 2 x 2 mm, with
to Wolfgang Schmitt, senior manager a height of 0.65 mm. The
of strategic marketing, end users have sensors can also be used for
an increasing demand for better, more advanced gesture recogni-
reliable, and more convincing use cases. tion, such as tilting of the
“That can be influenced generally by wrist, and can detect activ-
sensor performance,” he said, referring ities like running, walking,
specifically to higher precision. and standing still.
IMAGE: (FIG. 1) BOSCH SENSORTEC; (FIG. 2) ADEUNIS RF

This applies to six-axis inertial mea-


surement units (IMUs) that are being LoRaWAN and Sigfox
built into smartphones. Also, in fitness compete for best Fig. 2: The LoRaWAN sensor transceiver takes any sensor
trackers, “higher performance of the ac- wireless connectivity output and converts it to an RF signal for LoRaWAN wireless
celeration sensors and more sophisticat- As the IoT continues to connectivity.
ed step-counter algorithms directly result expand, the connection of
in more accurate step-counting figures.” sensors through low-power, wide-area While it’s tempting to integrate more
Two of Bosch Sensortec’s most recent networks over ranges of up to 50 km is intelligence into each new sensor, includ-
announcements, the BMA456 and now possible through LoRaWAN and ing wireless connectivity, that does add
BMA423, play directly to this theme, Sigfox. While range is critical for fast cost and space, while also requiring a
along with the trend toward adding deployment of hundreds or thousands sensor vendor to commit to one or more
more integrated functionality as well as of nodes with minimal concern about interfaces. This can result in multiple SKUs
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
24 Product Trends
Insight into current product developments

and potential inventory management devices: the AWR1642,


costs. There’s also the problem of what to AWR1443, and AWR1243
do about legacy sensors: How do they get (Fig. 3). They vary by
connected to the new wireless networks? level of ARM Cortex-R4
Adeunis RF’s transceiver abstracts the processor and DSP
wireless interface from the sensor by tak- support, as well as the
ing any sensor I/O signal and converting number of transmit and
it to operate over a LoRaWAN wireless receive channels and
network. The I/O can include 0–10 V their sampling rate (12.5
and 4–20 mA, among others. The result or 37.5 Msamples/s). All
is that both new and legacy sensors — operate in the 76- to 81-
such as those for temperature, pressure, GHz band.
humidity, and CO2, just to mention a few Sticking with automo-
— can be quickly and easily connected to tive, but on the transducer Fig. 4: The PGA460 is an AEC-Q100-qualified automotive
the IoT over LoRaWAN. side, TI also introduced the ultrasonic signal processor and transducer driver.
The interface has an RF power of 14 PGA460 ultrasonic SoC
dBm (25 mW), an RF sensitivity of up to with transducer driver and signal condi- Haptic-feedback control combines
–140 dBm, and a range of 10 km. It sup- tioner. The chip has a complimentary low- sensors and transducers
ports LoRaWAN V1 network protocol. side driver pair that can drive an ultrasonic There are many ways to interface with a
transducer either in a transformer-based machine or an automotive dashboard,
Automotive radar jumps to wider topology using a step-up transformer or but Ultrahaptics, a U.K. startup, has
bandwidth in a direct-drive topology using external developed a means of sensing a hand’s
The recent regulatory required shift high-side FETs (Fig. 4). It can also receive position and then sending precisely tar-
from 24 GHz to the 76- to 81-GHz band and condition the reflected echo signal for geted ultrasonic pulses to the fingers or
for automotive radar came with a num- object detection. hand surface, such that it can feel for all
ber of distinct advantages. These advan- Both radar and ultrasonic signals are the world like a user is touching a knob
tages include smaller-sized components being incorporated into semi-autono- or button in mid-air (Fig. 5).
due to the higher operating frequency, mous and autonomous vehicles (see Page
as well as wider bandwidth, the latter of 14, “Smart vehicles launch the autono-
which opens the door to higher-resolu- mous moonshot”). Along with LiDAR,
tion detection. cameras, and even infrared sensors, they
To help designers take advantage of are all being recruited into an optimal
the higher-band operation, Texas Instru- sensing system for maximum road safety
ments introduced three new radar sensor and driving effectiveness.

Fig. 5: Ultrahaptics’ haptic feedback system


can use any gesture-sensing technology. Its
differentiation is its use of ultrasonic pulses
IMAGE: (FIG. 3 & 4) TEXAS INSTRUMENTS; (FIG. 5) ULTRAHAPTICS

to the hand to provide 3D touch sensations.

Though the technology has been in


various stages of development since
2013, it is now gaining traction in
automotive and medical applications.
These two applications stand out
as a designer can use Ultrahaptics’
technology to develop a system that
doesn’t require visual coordination,
which is distracting for drivers. Nor
does it require doctors or others in the
Fig. 3: The AWR1642 is one of three 76- to 81-GHz corner radar solutions for automotive: It’s medical field to touch surfaces that
more focused on short-range applications than the other two. may be contaminated. ☐
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
No place to PROVEN

replace a battery. 40
YEAR
OPERATING

LIFE *

Highly remote locations call for Tadiran batteries.

Battery replacement is costly and often ANNUAL SELF-DISCHARGE


dangerous work. Reduce the risk with Tadiran
TADIRAN COMPETITORS
bobbin-type lithium thionyl chloride (LiSOCl2)
batteries. With an annual self-discharge rate of
just 0.7% per year, Tadiran LiSOCl2 batteries
enable low power consuming wireless devices to
operate for up to 40 years on a single battery, 0.7%
up to 4 times longer than the competition.
Our batteries also feature the highest capacity,
highest energy density, and widest temperature
range of any lithium cell, plus a glass-to-
metal hermetic seal for added ruggedness and
reliability in extreme environments.
Up to 3%
Tadiran Batteries
Take no chances. Take Tadiran batteries
2001 Marcus Ave.
that last a lifetime. Suite 125E
Lake Success,
NY 11042
1-800-537-1368
516-621-4980
* 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. www.tadiranbat.com
26 New Products
Packaging & Interconnections
Microwave, RF assemblies qualify stringent data rates are more sensitive to signal integrity issues during
routing and installation.
aircraft qualifications Gore claims that its microwave and RF assemblies, high-da-
Gore & Associates Inc. has announced ta-rate cables, high-power delivery cables, aircraft sealants, and
that its GORE-FLIGHT Microwave shielding materials are significantly smaller and lighter-weight.
Assemblies and GORE Microwave/ So they are more efficient at higher data speeds and more dura-
RF Assemblies have passed stringent ble for long-lasting performance.
U.S. Army flight qualifications on the W.L. Gore & Associates: www.gore.com
MH-47 Rotorcraft.
The 6 Series of GORE-FLIGHT
Microwave Assemblies claim to have a very low insertion loss before Cat5e Ethernet cables designed
and after installation. The lightweight yet rugged construction also for outdoor, industrial
improves fuel efficiency and increases payload.
GORE Microwave/RF Assemblies are also lightweight and
environments
easy to route in small spaces. Therefore, these assemblies ensure L-com has released two families
long-lasting electrical performance after installation and, thus, of outdoor-rated, high-flex Cat5e
lower the overall cost. cables for industrial Ethernet appli-
Gore has provided a complex harness, which includes more cations. These Cat5e patch cables are built to withstand harsh
industrial environments and repetitive motion applications.
than 100 of its rugged assemblies, for the ALQ-211 SIRFC, a
The TRD855HFT series cables feature a thermoplastic
critical aircraft survivability system (ASE) used by the 160th
elastomer (TPE) cable jacket that is resistant to industrial oils,
Special Operations Aviation Regiment (SOAR). ASE is a crucial
UV damage, and weld spatter while maintaining a CMX flame
requirement for avionic systems operating in harsh conditions.
rating. Next, the TRD855HFP series cables employ an econom-
Gore is also working with the U.S. Army to provide high-speed
ical polyurethane (PUR) cable jacket that is resistant to water as
data interconnects for the future APR-39 D (V) 2 radar warning
well as UV damage.
receivers on the Apache D model. Brian Tallman, product manag-
“These Cat5e cables are designed for industrial and outdoor
er for Gore’s North American Aerospace Team, noted that higher
environments in which Ethernet cables typically fail due to
high-stress conditions,” said Dustin Guttadauro, product man-
ager at L-com.
420°F, 720 VPM! Both TRD855HFT and TRD855HFP cables offer resistance
against sunlight, oil, and chemicals commonly found in the
High Temperature - High Dielectric industrial settings. According to Guttadauro, apart from eight
NOMEX® for Electrical/Electronic standard off-the-shelf lengths, these cables are also available at
Tubing, Bobbins, Fabricated Parts custom lengths to suit specific installation needs.
Widely specified for its excellent electrical, L-com: www.l-com.com
thermal and mechanical properties…
“Nomex” tubing is produced by Precision
with an exclusive, special binder to
provide a thermally Wet-mate connector qualified for subsea
stable class “H”
insulating tubing. control systems
U/L Recognized. TE Connectivity has unveiled
a HydraElectric wet-mate
connector that features a mod-
ular construction and enables
multiple configurations to be
assembled from a minimum
number of components.
The connector design
features an industry-standard
To receive literature & details fast - www.pptube.com panel, bulkhead, and stab plate interfaces allowing users to
Phone: 847-537-4250 • Fax: 847-537-5777 • E-Mail: [email protected] exchange, upgrade, or modify the existing hardware with min-
More Than 80 Years - The Original
imum effort. Moreover, it offers corrosion resistance without
cathodic protection.
The Seacon wet-mate connector has been designed, developed,
Paper Tube Company
1033 S. Noel Ave., Wheeling, IL 60090
tested, and fully qualified in accordance with the latest industry
™ ® DuPont Co. standards for subsea control system power distribution, includ-
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
New Products 27
Packaging & Interconnections
ing ISO 13628-6 and API 17F SEAFOM. And it’s available in remotely operated vehicle
(ROV), stab plate, and manual modes.
The connector’s modular construction comes with four, seven, or 12 electrical circuits,
and that enables multiple configurations for use in electrical, signal, and distribution
networks in subsea control systems. Bjørn Hallvard Lund, product manager of Aerospace,
Defense & Marine at TE Connectivity, says that the Seacon wet-mate connector comple-
ments TE’s full product offering for the subsea oil and gas market.
The rugged marine systems require connectors with high-speed communication
capabilities and the ability to manage a range of operating currents. TE’s new Hy-
draElectric connector supports copper-based high-speed Ethernet protocols, includ-
ing CAN bus communications.
They are qualified for both 1,500 Vdc and 1,000 Vac pin to pin. Furthermore, the
connector includes oil-filled dual barriers to help maintain pressure balance. And it
can be verified for use at 4,000-m water depth.
TE Connectivity: www.te.com

PCB tool for rapid prototyping of surface-mount ICs


ICStripBoard is a new tool that facilitates rapid prototyping of surface-mount ICs
available in a variety of packages and different sizes. ICStripBoard has been designed
to accommodate the majority of surface-mount ICs.
The surface-mount ICs are available in
four standard pitches of 0.5, 0.65, 0.95, and
1.27 mm. Pitch is the space between IC pins.
The prototyping tool has been designed as
long strips on thin 0.8-mm FR4 boards,
which can easily be cut to the correct amount
of pins that the IC is using.
That allows the strip to be cut for multiple
ICs on multiple projects. Moreover, the cut
pieces can easily be soldered and glued to other prototyping products. So the prototyping
board allows designers to experiment with multiple ICs without having to build a PCB.
It’s a far cheaper solution than buying alternative break-out boards simply because
engineers can cut them to size and the pattern repeats down the strip, allowing this to
be done multiple times.
ICStripBoard: www.icstripboard.co.uk

Optical interconnect demos PCIe links over FPGA board


Samtec has demonstrated its FireFly micro-flyover system for the Xilinx Virtex UltraS-
cale+ FPGA evaluation kit. In this demo, the 28G
x4 FireFly optical connector system transports four
28-Gbps transceiver signals generated by an FPGA.
These four 28-Gbps signals fly over the board
through an FQSFP twinax cable assembly to the
port and then loop back over the same path, where
they are received by the FPGA. Thus, the demo showcases the 28-Gbps performance
on all four links with zero-bit errors while having traveled over the twinax cable.
The FireFly mid-board technology supports the PCIe Gen 3 interface for low-la-
tency applications in data center environments. Samtec claims that its FireFly PCIe-
over-fiber solution allows storage and the GPU to be located as far as 100 m away.
The FireFly interconnect system enables chip-to-chip, board-to-board, on-board, and
system-to-system connectivity at data rates up to 28 Gbps. Moreover, its micro-footprint for
optical and copper interconnects is interchangeable within the same connector system.
Samtec: www.samtec.com
ELECTRONIC PRODUCTS • electronicproducts.com • AUGUST 2017
Product Mart 28 New Products
Electronic Products Presented by the Manufacturer Power Sources

40 to 300 ps Rise Time Pulser s 150-W AC/DC ref


Avtech offers over 35 models which provide 40-300 ps rise times design delivers 21 W/
with 5 to 100 Volt output amplitude and PRF to 200 MHz. These
models are ideal for testing high speed semiconductors and
in.3 and 95% efficiency
optoelectronics in both laboratory and factory floor applications. The NVE021A is a 150-W AC/DC reference design using the
Customized models are also available.
company’s GaN power ICs and offers >21 W/in.3 and >95% effi-
For datasheets, ciency. It is more than two times smaller than typical commercial
test results & pricing:
designs and 40% smaller than the previous best-in-class.
Avtech The design can be adapted over a power range of 85 to 500
Tel: 888-670-8729
Fax: 800-561-1970 W to enable a new class of converters to address ultra-thin LED
[email protected] TVs, fast-charging laptop adapters, high-density gaming sys-
http://www.avtechpulse.com/
50 ps/DIV, 5V/DIV tems, all-in-one PCs, and any other systems seeking high-den-
Model AVP-3SA-C
sity or high-efficiency power solutions.
The GaN power ICs enable high-frequency switching to
shrink passive components. Soft-switching critical-conduction
Turbo® Miniature Piezoelectric Alarms mode (CrCM) PFC and LLC DC/DC stages operate at work-
Tiny (approx.1”x1”) piezoelectric alarms provide LOUD output
with a super-sleek design. Available in variety of voltages and ing frequencies up to 300 kHz (600 kHz during start-up, burst
output up to 103dB! Rugged, tamper-proof, lo-profile panel mode). The reference design is available now for $465 each,
mount design is IP68 and NEMA 4X. Optional manual volume
control offers increased attenuation. ISO 9001:2000 registered which includes a user guide/test report with all schematic and
company – all products made in the USA. layout design files, plus bill of materials.
Floyd Bell Inc. Navitas Semiconductor: www.navitassemi.com
Tel: (614) 294-4000
Fax: (614) 291-0823
[email protected]
www.floydbell.com 600-W programmable supplies feature
remote sense without wires
Versatile Power introduces its BENCH XR line of 600-W
programmable supplies. They are the industry’s first to provide
accurate point-of-load
voltage regulation over a
wide range of load con-
ditions without the use of
an auxiliary remote-sense
circuit. Remote sense is
implemented using the
company’s patent-pending digital control to instantaneously ad-
just the voltage at the output terminals in response to changes in
output current. This technique eliminates output noise that is of-
ten introduced by conventional remote-sense wiring. In addition,
by eliminating the remote-sense wiring, the BENCH XR is not
susceptible to failures caused by ground loops or reverse-polarity
sense-wiring connections.
The BENCH XR series of 600-W programmable power
supplies are packaged in a 1U, ½-rack-width enclosure. It offers
Electrical Insulating Bobbins an 85- to 265-Vac input. Five models provide for output voltage
Fabricated or molded, hi-dielectric, hi-temp, choice of 12
materials, extra strong, with or without tie-offs, low cost, and current ranging from 30 V to 400 V and 2.5 A to 33 A. Digital
over 80 years of experience. Made In USA! display of output current and point-of-load voltage is provided.
To receive literature & details fast: Multiple standard I/O interfaces (Ethernet, USB 2.0, and ana-
www.pptube.com
log) facilitate setup and operation of test protocols. It provides
Precision Paper Tube Company command compatibility for Keysight 603x, Sorenson DLM, and
Phone: 847-537-4250
Fax: 847-537-5777
Xantrex XFR. The BENCH XR series is compatible with LXI Class
[email protected] C specifications and supports the Standard Commands for Pro-
www.pptube.com grammable Instruments (SCPI) language. The ability to perform
More Than 80 Years - The Original
remote sense without any added wiring both simplifies setup and
improves power supply performance and reliability.
Versatile Power: www.versatilepower.com
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS
Hioki LCR/Impedence Ad.qxp_IM7580-EP 5/9/17 9:42 AM Page 1

Great Performance & Price


Without Compromising Quality

IM7585 IM9200
Impedence Analyzer SMD Test Fixture

Additional models available

• Power measurement frequencies from 100kHz to 3GHz


• Multi Layer Chip Capacitor
• Mass Production/Product Development
• Inductors & Coils
• Transformers
• Piezo Electric Elements
IM3533-01 • RFID
LCR Meter

LCR & IMPEDENCE ANALYZERS


Over 80 Years of Meeting Your Needs!
ISO9001
ISO14001

 #        #     #  #       # 

   #  ! "   ! "  "  ! " #  #   #  ! "   ! "  "  ! 

w w w. h i o k i u s a . c o m
AUGUST 2017

c klin g Da ta Ce nt e
T a r
ER G Y CO NS UM PT IO
EN efficienc y w
N
hile reducing costs and emiss
ions p
asing age 2
Incre

Also inside
Selecting the optimum EMI test considerations Power path management 16th Annual Power
MOSFET for off-line switching for power supplies p8 is needed for Source Guide p13
power supplies p5 mobile applications p11
Power Supplement Contents 1

FEATURES
Power technologies that tackle data center
energy consumption
4 Data centers need advanced cooling strategies, hardware
virtualization, and direct-conversion architectures . . . . . . . . . . . . . . . . 2

Selecting the optimum MOSFET for off-line switching


power supplies
Get an understanding of the construction of the MOSFET and how
its interdependent parameters affect the power supply’s performance . .5

EMI test considerations for power supplies


5
Case studies show that conducted emissions tests should be done at
different voltages and loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Power path management for mobile applications


The DPPM control system can obtain power as soon as the input
source is applied, even if the battery is depleted . . . . . . . . . . . . . . . . . . .11

8 Source Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
View the full Source Guide at www. electronicproducts.com

12
COVER IMAGE: SHUTTERSTOCK; DESIGN: GIULIA FINI

ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017


2 Power Supplement

Power technologies that


tackle data center energy
consumption
mediate bus converters) down-convert
Data centers need advanced cooling strategies, from the 48-V bus to a typical voltage
hardware virtualization, and direct-conversion level of 12 V. The establishment of the in-
termediate bus was primarily designed to
architectures save the cost and bulk of multiple isolat-
tric-based solutions, to deliver supplies of ed DC/DC converters. Beyond the IBCs
BY BOB CANTRELL
Senior Application Engineer, low-carbon electricity. Of course, both of are the onboard and non-isolated PoL
Ericsson Power Modules these approaches can be used together to converters, which further step down the
www.ericsson.com/powermodules reduce energy consumption. 12-Vdc intermediate bus to the voltages

E
required by board-level components such
nergy consumed by data centers Power architecture today as processors, ASICs, FPGAs, memory
around the world has increased The second overall area of contribution ICs, and other logic devices. Voltages can
dramatically over the past decade or to data center energy consumption is ob- typically range below 1 V as needed to
so, going from almost nothing to over 2% viously the power required for operation power processors or FPGA core logic.
of the global electricity supply. Added to by servers, mass storage systems, and The deployment of two down-con-
this is the production of approximately 2% data communications and networking version stages has the advantage of
of worldwide greenhouse gas emissions. equipment. In this area, new and innova- delivering optimal balance between
Predictions exist today that data center tive modern power techniques are being the intermediate bus supplying the
energy consumption is to triple over the sought within the industry to reduce ener- PoL converters and the load currents
next decade, starting to rival and even supplied by the PoLs, which is import-
overtake the electricity consumed by entire ant for maximizing power-conversion
Data center energy
countries with moderately large economies. efficiency at the system level. The 12-
consumption has increased Vdc voltage level was chosen to ensure
Cooling concerns dramatically over the past high enough voltage to deliver all the
One important area of contribution to power required by the load, or board,
energy consumption is delivering the nec- decade in part because in times of high network data traffic,
essary cooling for equipment using mas- of the power required for in addition to delivering lower distri-
sive air-conditioning units and systems. bution losses. However, the approach
operation by servers, mass can lead to low efficiencies, especially
There are a couple of ongoing approaches
to reduce consumption in this area. The storage systems, and so when there is low traffic demand.
first, which is one being pursued by many data communications and An overview of the intermediate bus
major data center providers, is the raising architecture is shown in Fig. 1.
of the maximum ambient temperature networking equipment. However, the overall efficiency of
marginally above typically accepted limits this entire network from the AC line
for the operation of servers and other gy consumption in modern data centers. is approximately 85%, which means
equipment, but not enough to introduce Historically, going back to the 1980s at up to 15% of the power to the system
significant reliability issues. A second least, in telecoms and datacoms applica- is, therefore, dissipated as heat by the
approach is to benefit from climates with tions, the distributed power architecture various power devices. This heat needs
significantly lower average temperatures has been dominant, in which an AC line to be removed to enable high system
and locate data centers in cooler parts of voltage is rectified to typically deliver 48 reliability, or at least within well-defined
the world, such as in the Scandinavian/ Vdc for distribution to server cabinets via acceptable limits.
Nordic countries: Major companies such individual onboard and tightly regulated
as Facebook are already establishing facili- point-of-load (PoL) converters. Trends
ties in Denmark and Sweden, for example. However, this architecture has evolved There are a number of ongoing trends
In addition, these countries, among oth- further over the past decade or so with in the industry that seek to maximize
ers, are increasingly looking to renewable the introduction of the intermediate bus efficiency, reduce power consumption,
sources of energy, including hydroelec- architecture (IBA), in which IBCs (inter- and minimize costs. There is the move
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT
Power Supplement 3

toward advanced software-based archi- Those who have been early adopters over a wide range of transients, which
tectures, such as software-defined net- of digital power, including those set- is a major time-saver and can signifi-
working (SDN) and network function ting up data center facilities, have seen cantly enhance time-to-market. The
virtualization (NFV); the increasing use the many advantages of its enhanced tools are also available to apply track-
of digital power and the extension of flexibility. For example, the simulation ing and sequencing of power rails that
which is the emergence of the soft- capability of the Ericsson Power De- have multiple loads for a given piece
ware-defined power architecture; and signer software can quickly determine of silicon. In addition, output voltages
the re-emergence of direct-conversion the optimized number of capacitors can be changed on the fly, as well as the
technology, which has actually been used on each rail to ensure stability ability to monitor voltages, currents,
around for 35 years or so, to down-con-
vert the traditional 48-Vdc line directly
to the board-level voltages required by
processors and other logic ICs.
Addressing the first of these: SDN,
essentially, is about separating the con-
trol plane from the data plane — or the
software from the hardware. Application
software does not necessarily need to
run on specific networking hardware,
but perhaps on servers in a data cen-
ter. In addition, NFV offers significant
economies of scale and standardization:
Multiple functions or applications can be
consolidated on “virtual machines” based
on commercially available hardware. Vir-
tualization can certainly make a contri-
bution to reducing energy consumption,
but software control at the board level is
also critical.

Digital power The Future of Circuit Protection for


The second element is digital power.
Generally, traditional analog switched- 24VDC Switch Mode Power Supplies
mode power converters reach peak
E-T-A’s new REX12-T electronic circuit protector
efficiency just below the maximum
transforms how 24VDC switch mode power supplies
load. While efficiency is only slightly
reduced at maximum load, it is signifi- are protected and offers engineers:
cantly lower at half-load or at times of • Time savings: innovative mounting &
low demand, for example. The reason
connection technology reduces installation time
for this is largely due to the use of
by 50%
fixed-value capacitors, which are em-
ployed to ensure a stable supply over • Space savings: two channels are only 12.5 mm
a wide range of operating conditions. wide
The downside is a lack of flexibility and
the inability to deliver high or uniform • Design exibility: modular design makes
efficiency across the potential range system upgrades easy
of loads. However, the possibilities
• Increased machine uptime: Smart version
offered by digital power can overcome
this limitation. The digital control loop (REX12D-T) provides monitoring and remote
can be easily and quickly changed and diagnostics with IO-Link connection
optimized via digital monitoring and
control. In addition, the use of digital
power converters requires fewer ex- See the REX12-T in action:
ternal components, thereby improving www.e-t-a.com/EP_REX12-T
reliability and saving board space.
ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017

PA-EP-2017-08-APC-REX12-Ad.indd 1 7/7/17 8:42 AM


4 Power Supplement

As mentioned previously, the


direct-conversion concept is not a
new one, but the reduction in size and
competitive pricing of direct-conver-
sion DC/DC modules is bringing the
architecture back to the forefront. The
most significant reason for this change
in approach is the simplification of
current distribution due to the increas-
ing power requirements of modern
data centers and the required reduction
in I2R distribution losses. Distributing
12 V at high current to some of the
Fig. 1: Voltage conversion stages in typical data center or datacoms facility.
leading-edge server cabinets used in
and temperatures. to adjust the output voltage to optimize data centers today is prohibitive and
In digital-power-based systems, performance of individual processors. not practical. However, direct-conver-
communication between IBCs and PoL sion architectures can distribute power
converters is carried out via the PM- Direct conversion at 48 V rather than 12 V. The bus sup-
Bus, which is an industry standard de- The third trend is direct conversion, plying the converter carries just 25%
veloped out of the SMBus and defines which is likely to offer significant ad- of the current that would be required
the physical connection and the data vantages to many data center operators. to deliver the same level of power at 12
exchange protocol between a central The principle involves bypassing the V, which also means a reduction in the
controller and power modules. use of IBCs and PoL converters and amount of copper bus bars and cables.
deploying a new and highly efficient
Software-defined power type of power module that converts Conclusion
Further extending the digital power the 48-V bus voltage typically used in Energy consumption is a major issue
trend is an important next evolution- datacoms applications directly down for data center operators. New and
ary step, which is the software-defined to the PoL board-level voltages, which advanced power architectures and
power architecture (SPDA). This intro- can be below 1 V, but in a single power techniques will be required to increase
duces real-time adaptability to digital stage (see Fig. 2). system efficiency and reduce costs as
power supplies and has the potential to The latest IBCs typically offer effi- well as carbon emissions. In addition
bring truly energy-efficient and pow- ciencies of 95–96%; and PoL converters to implementing advanced cooling
er-optimized board-level capabilities typically can offer more than 90%. strategies to deal with heat dissipation,
into advanced network applications. Combined energy loss, however, can important technologies will include
In the SPDA, advanced processors are reduce overall efficiency to approxi- hardware virtualization, digital power
able to use software command control mately 86.4%, whereas a single con- and its SDPA extension, and in many
to adjust output voltages to increase verter can deliver efficiencies of 89% or potential applications, the deployment
processor performance or lower voltag- more for the same given load. of direct-conversion architectures. ☐
es at times of low load demand.
The SPDA is essentially a scheme
that implements many advanced
techniques, including concepts such
as the dynamic bus voltage (DBV) and
adaptive voltage scaling (AVS), as well
as others, including fragmented power
distribution or phase spreading. An
evolution of the IBA, the DBV dy-
namically adjusts the intermediate bus
voltage of the IBC as a function of load
current to deliver higher power conver-
sion efficiency. AVS is a technique to
optimize supply voltages and mini-
mize energy consumption in modern
high-performance microprocessors. It
uses a real-time closed-loop approach Fig. 2: Conventional two-stage IBC + POL converter versus a direct-conversion solution.
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT
Power Supplement 5

Selecting the optimum


MOSFET for off-line switching
power supplies
Get an understanding of the construction of the MOSFET and how its
interdependent parameters affect the power supply’s performance
BY MARTY BROWN (Fig. 1b) uses its die depth to increase the channel area, which
Senior Consultant, D3 Semiconductor significantly reduces its RDS(on) for a given die area. By greatly
www.d3semi.com increasing the conduction area within a much smaller surface

T
area, superjunction MOSFETs allow for a smaller die size for
oday, the choice of silicon MOSFETs falls into two each current rating as compared to the planar MOSFET. As
major divisions: the planar and/or trench MOSFET an added benefit, they offer much lower parasitic capacitances
and the superjunction MOSFET. Planar MOSFETs with much faster switching speeds..
have their active region close to the surface of the die (Fig. Table 1 compares the key parameters between a standard
1a). To decrease the RDS(on), additional die area is required, planar MOSFET and a high-performance +FET superjunction
and to increase the voltage of the MOSFET, the die must be MOSFET, newly introduced by D3 Semiconductor, which
made thicker. In comparison, the superjunction MOSFET have nearly identical drain voltage and current ratings. As
shown, the RDS(on) and the MOSFET’s annoying parasitic
capacitances are much lower with the superjunction MOSFET,

By greatly increasing the conduction


area within a much smaller surface
area, superjunction MOSFETs allow
for a smaller die size for each
current rating as compared to
planar MOSFETs.

and the switching speed is more than three times faster. The
Fig. 1a: Standard planar MOSFET construction. truly remarkable factor is that the uncapped die area of the su-
perjunction MOSFET is one-third that of the planar MOSFET.
The efficiency comparisons between a superjunction MOS-
FET and standard planar MOSFET speak volumes about how
a designer can improve cost and switching performance of a
typical off-line switching power supply. D3 Semi’s superjunc-
tion MOSFET exhibits a more-than-73% reduction in its die
area, a more-than-50% reduction in input gate capacitance
(Ciss), and a more-than-69% reduction of the value of the
miller capacitance (Crss).

MOSFET performance in power supplies


Fig. 2 illustrates a very common plot when testing the effi-
ciency of a switching power supply over its expected load
range. The efficiency exhibits a peak at some load between the
maximum and minimum loads. With a good design, that peak
Fig. 1b: Construction of a generalized superjunction MOSFET. will be at or near where the supply operates more than 90%
ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017
6 Power Supplement

That peak of the efficiency,


as seen in Fig. 3, can help with
selecting an “optimum” MOS-
FET for your design and oper-
ating range. That is where the
peak efficiency occurs where
your supply spends more than
90% of its operating life.
Fig. 4 is a qualitative
representation of the behav-
iors of the three losses over
the load range. This only
Table 1: Comparison of a representative standard versus a D3 describes the losses indi-
superjunction MOSFET. vidually and not collectively. When these losses are combined
(summed), the result is much more informative.
This example shows what can happen when the design-
er chooses the lowest RDS(on) MOSFET within a family of
high-voltage MOSFET offerings. The result is an overall lower
supply efficiency as compared to a higher RDS(on) (smaller die
area) device being selected. (Of course, there is a whole set of
variations in the results given the range of MOSFET parame-
ters and operating conditions. This represents a worst possible
case outcome of a poor design decision.)
Choosing the lowest available RDS(on) MOSFET increases
the Ciss and Crss, which, in turn, requires a higher peak gate
drive current. Since the gate driver cannot supply that peak
current in the short period, the MOSFET switches slower
Fig. 2: Efficiency versus load — the influences. (longer tSW and lower dVDS/dt). This drastically increases
the significant drain-source switching loss and the gate driver
of its operating life. Its location and its peak value is totally loss. The gate driver experiences a larger loss related to the
determined by trading off the various MOSFET parameters amplitude of the peak gate current and the lengthened time
during the design process. required to switch the MOSFET.
At light loads, the on-time (duty cycle) is small. This makes
the MOSFET conduction losses much less significant. Here,
the switching and gate drive losses dominate its efficiency.
To get a very clear understanding of
Within the typical power supply, the switching frequency is the performance of the power supply
relatively constant, so the switching and drive losses are also
relatively constant. over its complete operating range,
At higher loads, the MOSFET (and output rectifier) con- one should consider the variation in
duction losses become more significant than the drive and
switching losses. Here, the value of RDS(on) becomes important. efficiency over the expected range of
As with all resistive switches, the supply’s efficiency deterio-
rates as the output load increases. This non-linear effect is the
output load and input voltage.
ID2(RDS(on)) loss.
The resulting drive and drain-source switching loss can
Design considerations make the efficiency noticeably worse at the lighter loads and
There are many subtle aspects when considering the best also arithmetically lowers the combined efficiency curve by
MOSFET for a switching power supply design. Conventional the amount of these losses. This degrades the overall efficien-
wisdom holds that the lowest RDS(on) yields the most efficient cy, especially at the lighter loads, and pushes the peak efficien-
supply. However, looking only at a power supply’s efficiency at cy point toward the higher loads.
one operating condition can be very misleading. To get a very The perceived benefit of a reduced RDS(on) does, indeed,
clear understanding of the performance of the power supply decrease the conduction loss, but its overall effect is somewhat
of its complete operating range, one should consider the vari- marginal. The conduction loss is mainly responsible for the
ation in efficiency over the expected range of output load (W) “curviness” of the efficiency curve. The degree of curvature
and input voltage. becomes lesser for lower RDS(on)s. The increase in the efficiency
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT
Power Supplement 7

Fig. 3: Example of choosing the lowest RDS(on) MOSFET within a Fig. 4: The behavior of the three losses over the load range.
supply.
made once and the results are accepted. To the informed
due to the reduced RDS(on) is only arithmetic. The conduction designer, this is only acceptable when the MOSFET behavior
loss is still overwhelmingly dominated by the square of the is more thoroughly examined.
drain current. This helps at the higher loads, where the effi- There are many subtle aspects when considering the best
ciency is dominated by the conduction loss. In the case of Fig. MOSFET for a switching power supply design. A little ele-
4, the increase in the switching and drive losses were larger mentary understanding of the construction of the MOSFET,
than the improvement in the raised efficiency level and the as well as how its interdependent parameters affect the power
curvature. This is a very possible outcome. supply’s performance, allow the designer to see why migrating
For the average engineer, not really understanding the from standard planar MOSFETs to superjunction MOSFETs is
physical influences of MOSFETs, the choice of MOSFET is compelling. ☐

ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017


8 Power Supplement

EMI test considerations


for power
supplies
Case studies show that
conducted emissions
tests should be
done at different
voltages and loads
BY LORENZO CIVIDINO
Director, Global Applications
and Support;
HAI HO
Technical Applications Support
Engineer, SL Power Electronics
www.slpower.com

W
hen selecting a power supply, Fig. 1: An OEM’s 250-W external power supply failing emissions at low load conducted
one of the most important emission at 25-W load (10% of rated load).
functional requirements Blue: Quasi-peak scan. Green: Average scan.
of the product is its EMI performance. BQ line: CISPR11 class B l Quasi-peak limit. BA line: CISPR11 class B Average limit.
AC/DC power supplies are the interface
between the end-product electronics and
the power utility providing the energy
to power the end application. FCC and
International Standards require product
EMI performance to comply with the
emission limits defined in the standards
depending on the application.
A common assumption is that
EMI emissions are highest at full-load
conditions or at a certain AC line
voltage range. However, depending on
the power conversion topology of the
power supply, the emissions can vary
significantly over line voltage and over
its load range. This article discusses
considerations when performing verifi-
cation tests and provides data showing
significant EMI performance variations
over load ranges. In some cases, low
load can be significantly worse than
full load. Recommendations are made
on the load and line conditions to Fig. 2: An OEM 30-W Level VI external power supply.
perform the verification testing. These Blue: Quasi-peak scan. Green: Average scan.
recommendations ensure proper and BQ line: CISPR11 class B Quasi-peak limit. BA line: CISPR11 class B Average limit.
thorough evaluation while also mini- Note: Green plot (Ave. scan) exceeds Average limit line.
mizing the tests required to achieve a Conducted emission at 240-Vac 30-W load (100% of rated load).
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT
Power Supplement 9

high confidence level of compliance. To illustrate the issue, the following are standards. These operating conditions
The endless pursuit for higher effi- some actual examples: have changed the way we look at the
ciency is driven by customers’ desires for A common technique uses a variable conducted emissions.
smaller, lower-cost power supplies. In switching frequency to reduce power Case Study #1: Power supply topology:
addition, government regulations (DoE loss across line voltage and load current, Resonant half-bridge with front-end
Level VI, EU ErP, etc.) are mandating also the burst mode/phase manage- PFC. Interleaved transition-mode PFC
higher-efficiency adapters, which have ment (pulse skipping or turning off one controller followed by a DC/DC stage LLC
motivated the industry to find innovative phase at light load), which is extremely variable-frequency resonant converter
solutions to meet these requirements. useful to meet light load efficiency with burst mode.
For these reasons, novel solutions have
surfaced with some unintended conse-
quences to EMI performance.
External power supplies use variable
modes of operation to reduce no-load
power and achieve high efficiency. For AIA
CONVERTER
example, as the output load decreased,
the converter may switch to a burst
mode of operation in order to minimize
power consumption. This can result in Next-Gen Platform
resonance of the EMI filter and cause an
increase in the conducted emissions at
of DC-DC Converters
for Military & High Reliability Applications
these lower loads compared to high- • -40 to +105°C Operation (optional -55°C) • High Power Density / Compact Size
er-load conditions. This situation may • Compliant with Military Transient Standards • No optocouplers for high reliability
be overlooked if there is an assumption • Integrated Soft Start and LC Filter • MIL-STD-461 Compliant with Filter
• Synchronization Circuitry • Encapsulated with Metallic Enclosure
that conducted emissions are worse at
full-load conditions.
Higher-power converters (~80+ W),
8 Watts: MGDD-08 Series
especially those needing to meet class C
27.5mm
27.5mm / 1.083”
/ 1.083”

• Ultra Wide input ranges


or D harmonic current, will have a pow- - 4.5-33VIN Range (45V ≤ 100ms transient)
V Trim
V Trim
Sync / 1.083”-VOUT-V1OUT 1
27.5mm
Sync
19.3mm
19.3mm
er factor correction (PFC) stage to mini- - 9-60VIN Range (80V ≤ 1sec transient) 0.76”
0.76”
UVLO UVLO
Set Set
-VIN -VIN
+VOUT 1OUT 1
V -V
Trim
+V
-V2OUT 2
• Dual isolated / unbalanced outputs for 3.3 ~ 50VOUT
OUT

mize input current harmonics. There are • DO-160 & MIL-STD-704 compliant 19.3mm
InputInput
Sync
+VIN+V
UVLO
FilterFilter -V+V
SetIN
OUT 1 2OUT 2
+V
OUT
+VOUT 1

various methods of achieving the PFC.


0.76” -VIN
• MTBF >1.2M Hrs @ 40°C per MIL-HDBK-217F Height: 8.0mm-V/ OUT 2
0.315” Tall
Input Filter +VOUT 2
These include transition mode control, +VIN

two-phase PFC stage — in which one 32.7mm


32.7mm / 1.287”
/ 1.287”
20 Watts: MGDD-21 Series
phase may be disabled at lower loads, SyncSync
• Ultra Wide input ranges
then enabled and phase-shifted to - 4.5-33VIN Range (45V ≤ 100ms transient)
UVLO UVLO32.7mm
Set Set / 1.287”
-VIN -VIN
reduce input ripple current at higher - 9-60VIN Range (80V ≤ 1sec transient) 26.1mm
26.1mm Sync
InputInput
UVLO
FilterFilter
SetIN
V Trim
V Trim
1.03”
1.03” +VIN+V -VOUT-V1OUT 1
loads — and fixed and variable fre- • Dual isolated & unbalanced outputs for 3.3 ~ 50VOUT -VIN +V1OUT 1
+VOUT
• DO-160 & MIL-STD-704 compliant 26.1mm Input Filter V -V
Trim -V2OUT 2
quency modes, to name a few. All these • MTBF >1,060kHrs @ 40°C per MIL-HDBK-217F 1.03” +VIN -V+V
OUT

OUT 1 2OUT 2
+V
OUT
+VOUT 1
techniques offer advantages in reducing Height: 8.0mm / 0.315” -VOUTTall
2
57.9mm57.9mm / 2.28”+V 2
/ 2.28”
power losses over a wide AC mains vari- OUT

ation and wide load variation. However, 150 Watts: MGDS-155 Series -VIN -VIN 57.9mm / 2.28” -VOUT-VOUT

they also pose a challenge in designing • Ultra Wide input ranges 36.8mm
36.8mm
1.45”
1.45”
SyncSync Sense
Sense (-) (-)
V Trim
V Trim
- 4.5-45VIN Range (50V ≤ 100ms transient) -VOUT
an EMI filter that will be effective at
-VIN
On/Off
On/Off SensSens
(+) (+)
- 16-80VIN Range (100V ≤ 100ms transient) 36.8mm Sync
+VIN+VIN
Sense (-)
+V +VOUT
heavy and low loads with less damping. - 150-480VIN Range 1.45” V TrimOUT
On/Off Sens (+)
To compound the matter, system-level • MIL-STD-1275, MIL-STD-704 & DO-160 Compliant +VIN +VOUT
• Single outputs from 3.3 ~ 28VOUT
EMI filters are often added by system • MTBF >490kHrs @ 40°C per MIL-HDBK-217F
Height: 12.7mm / 0.50” Tall
design engineers to address EMI sources
generated by the system electronics that Visit our website for new detailed product specifications & application notes
may be powered by the AC/DC power
supply. In many cases, the added sys-
tem-level EMI filter can make matters
worse due to the impedance mismatch
of the system EMI filter and the AC/DC
power supply EMI filter.
w w w . g a i a - c o n v e r t e r. c o m
ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017
10 Power Supplement

Supplier A Supplier B

In this example, the power supply (See Fig. 2) Star Level VI external power condition. The result showed that it
failed 2.5 dB over the limit at 170 kHz supply with green-mode variable-fre- passed by 5 dB under the limit.
at 120 Vac while passed by 9.5 dB under quency fly-back topology. Supplier B worked well with the
the limit at 240 Vac. It turned out to be We tested the power supply at power supply under test.
at 120 Vac, and mid-load, the interleav- different line and load conditions. We
ing PFC circuit turned off one phase found that it passed conducted emis- Conclusion
and doubled the other phase current to sions with a comfortable margin at 120 In many end applications, peak power
boost the voltage up to 400 Vdc, thus Vac full-load but lost all of its margins occurs for a short period of time,
increasing the peak current through the at 240 Vac and actually failed at re- while the constant load power is often
inductor and generating more emis- duced load regardless of line voltage. It at around 50% or less of the rated
sions than at 240 Vac. The noise was also failed Average Conducted Emis- power of the supply. When in idle
picked up by the nearby input line filter sion at a 15-W load (50% of rated load) mode, the system may enter the low
and conducted to the line. If measured and 3 W (10% rated load), regardless of power consumption mode. The sys-
at full load, the interleaved converter line voltage. tem is required to comply with EMI
would be operating with both phases emissions levels through all of these
and have lower conducted emissions Examples of system filters negatively operating conditions. With the end
than operating at lower loads with a impacting EMI performance: Two dif- application in mind and lessons from
single phase. With some modifications, ferent power entry modules were used the above case studies, it is essential
a solution was implemented with a 6-dB with the same power supply to demon- to perform conducted emissions tests
margin under all line and load condi- strate its impact to the EMI. at 100 Vac, 120 Vac, and 240 Vac and
tions. The power supply under test with load current sweeps from zero to full
resistor load had the CE margin of at load, or at least at 100%, 50%, and
Case Study #2: A supplier’s off-the-shelf least 6 dB under the Class B limit. 10% of rated load to ensure compli-
250-W external power supply used a The power entry module from sup- ance to internal norms.
resonant LLC half-bridge converter with plier A was first used as the interface Power entry modules are designed
burst-mode operation at low loads with between mains and the power supply as a low-pass filter, controlling both the
continuous conduction mode fixed-fre- under test. The conducted emission common-mode and differential-mode
quency PFC front-end. was tested and found that it failed by current. When in the system, it could
The power supply failed by 2 dB over 3.5 dB over the limit. cause resonance due to its inductive
the limit at 10% load, but passed with 0.6 Supplier A filter does not work well and capacitive reactance interfering
dB at 50% load and 5 dB under the limit with the power supply under test. with the power supply EMI filter.
at 100% load, similar to the issue that we The power entry module from sup- Choosing an appropriate power entry
found in Case Study #1. plier A was replaced with a filter from module with the right capacitance
supplier B. The conducted emission and inductance to use with the power
Case Study #3: An OEM’s 30-W Energy was performed again with the same test supply is critical. ☐
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT
Power Supplement 11

Power path management is


needed for mobile applications
The most-often-used DPPM control system can obtain power as soon as
the input source is applied, even if the battery is depleted

BY AARON XU

Save A
Technical Marketing Manager
Monolithic Power Systems
www.monolithicpower.com

A Stack!
commonly used power
path management scheme,
dynamic power path man-
agement (DPPM), is a control loop

Go Direct
that adjusts the charge current dy-
namically based on the input source
current capability and load current
level to achieve a minimum charge
At MicroPower Direct we keep costs down and pass the savings on to
time for a given source and system our customers. Get great customer service, fast on-time deliveries and
load. With DPPM, the system can experienced technical help on well over 5,000 standard power conversion
obtain power immediately once the products. We continue to introduce new power products (like those below)
on a regular basis. All at very low cost. So, put us to the test; call today
and save a stack (or two).
In mobile devices with a
MD900xx Series
rechargeable battery, a • 9W Output Power
charger IC is needed to • 2:1 and 4:1 Input Ranges
• Single & Dual Outputs
charge the battery when • -40°C to +85°C Operation
an external power source • Miniature SIP Package

is applied. The system


ML100xSRI Series
load inside the mobile • 1W Output Power
device could be provided • Miniature SMT Package
• 4.2 kVDC Isolation
by the battery, the input • Single & Dual Outputs
source, or both. • Available on Tape/Reel

MPO-500 Series
input source is applied, even with
• 500W Output Power
a deeply discharged battery. The • Compact Open Frame Construction
system voltage regulation method is • EN 60950 Approved
also discussed. • > 0.95 Active PFC
In mobile devices with a re- • Meets EN 55022 B
chargeable battery, a charger IC is • Only 5.0” x 3.0” x 1.4”
needed to charge the battery when Don’t sacrice performance or service for price. Call MPD today for a surpris-
an external power source is applied. ingly fast & low cost quotation. For product datasheets or free samples, go to:

www.micropowerdirect.com
The system load inside the mobile
device could be provided by the
battery, the input source, or both,
depending on the connection of the MicroPower Direct
battery and system load. A power We Power Your Success - For Less!
path management scheme is needed
ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017
MPD EP1708.indd 1 6/27/2017 11:42:13 AM
12 Power Supplement

charge current increases, the battery


DC/DC FET obtains a stronger gate drive
Charger IC Rails and smaller on-state resistance (RDS)
VIN VSYS until the battery FET is complete-
or DC/DC Backlighting ly turned on. When the discharge
current goes lower, the ideal diode
Battery 3G Module loop generates a weaker gate drive
FET and larger RDS(ON) to keep a 20-mV
difference between the battery and
system until the battery FET is
Fig. 1: NVDC power path VBATT turned off.
management structure. VSYS regulation in DPPM mode
can be flexible, depending on the
system requirement. If the front-
to handle this kind of power source end converter from the input to the
selection. Dynamic power path system is an LDO, VSYS can be set at
Dynamic power path management a level to specially benefit the system
management is the most
is the most popular scheme for power requirement. For example, VSYS is
path management in mobile applica- popular scheme for power 4.65 V for the MP2661 and 5.0 V for
tions. The basic power stage structure path management in the MP2660.
for DPPM is shown in Fig. 1. If the front-end converter from
mobile applications. With
In the DPPM system, the system the input to the system is a DC/
load is connected to the system bus DPPM, the system can DC converter, VSYS is usually set to
(VSYS). VSYS can be powered from obtain power immediately follow the battery voltage to im-
the battery through the battery FET prove efficiency. This is commonly
once the input source is
or from the input source through a referred to as a narrow voltage DC
DC/DC converter or LDO. When applied, even with a deeply (NVDC).
the input source is not available, the discharged battery. There are several advantages for
battery FET is fully on, so the battery DPPM control. First, the system
provides power to the system load. gains power immediately once the
When the input source is applied, dropping. Once VSYS drops below input source is applied, regardless
VSYS is regulated by the input DC/DC the battery voltage (VBAT) level, of whether the battery is depleted
converter or LDO. Simultaneously, the battery provides power to VSYS or not. Second, the charge current
VSYS provides a charge current to the through the battery FET. This is is adjusted dynamically based on
battery through the battery FET. In called supplement mode. In supple- the input source and system load to
this charging mode, priority is given ment mode, the input source and achieve a minimum charge time.
to the system load, and the remain- battery provide power to the system The limitation for DPPM control
ing power is used for charging. The simultaneously. lies on the fact that it is complicated
charge current is adjusted dynami- Before entering supplement mode, to ensure a smooth transition be-
cally based on input source capability if the battery FET is in linear mode tween the different operation modes.
and system load level, achieving a (not fully on; for example, when VBAT Usually, a VSYS loop, ideal diode loop,
minimum charge time. < VSYS_MIN + DV, or during a start-up charge voltage, and charge current
During the above charging pro- transient), to ensure a smooth transi- loop are required for battery FET
cess, if the system load is over the tion in and out of supplement mode, control.
power capability of the input source, an ideal diode mode is preferred to With DPPM control, the system
VSYS will drop. Once VSYS drops to a control the battery FET, such as the can obtain power as soon as the input
DPPM threshold, the DPPM control one in the MP2624A. source is applied, even if the battery is
loop activates and reduces the charge During the ideal diode mode, depleted. The charger IC with DPPM
current automatically to prevent VSYS the battery FET operates as an ideal control can also optimize the charge
from dropping further. This process diode. When the system voltage is current to fully use the input source
is also called DPPM mode. 40 mV below the battery voltage, the current capability. Although the con-
In DPPM mode, if the charge battery FET turns on and regulates trol for DPPM is complicated, DPPM
current is reduced to zero, and the the gate driver of the battery FET. is widely used in charger ICs that re-
system load is still over the input The voltage drop (VDS) of the battery quire power source selection, such as
power capability, VSYS continues FET is about 20 mV. As the dis- with the MP2624A and MP2660. ☐
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT
Power Supplement Source Guide 13

DC/DC power supplies


AC/DC power supplies

Circuit protection

semiconductors
Cooling devices

Transformers
Capacitors
Batteries

Resistors
Discrete
devices
Company URL
Aavid Thermalloy www.aavidthermalloy.com •
Abracon, LLC www.abracon.com •
Absopulse Electronics, Ltd. www.absopulse.com • • •
Acopian Power Supplies www.acopian.com • •
Allegro MicroSystems, LLC www.allegromicro.com • • •
Alpha & Omega Semiconductor www.aosmd.com •
American Electrical, Inc. www.americanelectrical.com •
American Power Design, Inc. www.apowerdesign.com •
Americor Electronics, Ltd. www.americor-usa.com •
AMETEK Programmable Power, Inc. www.programmablepower.com • •
AMETEK Dynamic Fluid Solutions www.ametekdfs.com •
Anaren Microwave, Inc. www.anaren.com •
API Technologies Corp. www.apitech.com • • • •
Applied Power Systems, Inc. www.appliedps.com • • •
Astrodyne TDI www.astrodynetdi.com • •
Avago Technologies www.avagotech.com •
B & B Battery USA, Inc. www.bb-battery.com •
B + B SmartWorx www.bb-elec.com • •

CHECK OUR INVENTORY AT:

www.fusesunlimited.com
CALL: (800) 255-1919

The Circuit Protection Specialist


ISO 9001 Registered

ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017


14 Source Guide Power Supplement
Advertisers of this supplement are shown in red.

DC/DC power supplies


AC/DC power supplies

Circuit protection

semiconductors
Cooling devices

Transformers
Capacitors
Batteries

Resistors
Discrete
devices
Company URL
Behlman Electronics, Inc. www.behlman.com • •
Bel Fuse, Inc. www.belfuse.com • •
Beta Dyne, Inc. www.betadynepower.com • •
Bias Power, LLC www.biaspower.com •
Calex Manufacturing Co., Inc. www.calex.com • •
Carling Technologies, Inc. www.carlingtech.com •
Central Semiconductor Corp. www.centralsemi.com • •
Chroma ATE, Inc. www.chromaus.com •
Chroma Systems Solutions, Inc. www.chromausa.com •
CKE/HVCA www.cke.com • •
Coilcraft, Inc. www.coilcraft.com •
Coilcraft CPS (Critical Products and Services) www.coilcraft-cps.com •
Communication Coil www.communicationcoil.com •
Cornell Dubilier Electronics, Inc. www.cde.com •
Cosel U.S.A., Inc. www.coselusa.com • •
CTS Corp. www.ctscorp.com •
CUI, Inc. www.cui.com • •
Curtis Industries www.curtisind.com • • • •
Cymbet Corp. www.cymbet.com •
Datatronic Distribution, Inc. www.datatronics.com •
Dean Technology, Inc. www.deantechnology.com • • • • •
Eaton Corp. — Pulizzi Products www.powerquality.eaton.com • • •
EBG Resistors, LLC www.ebg-us.com •
Efficient Power Conversion Corp. www.epc-co.com •
Enargy Power Corp. www.enargypower.com • •
Endicott Research Group, Inc. www.ergpower.com • • •
EPCOS, Inc. www.epcos.com • •
Ericsson Power Modules AB www.ericsson.com/powermodules •
E-T-A Circuit Breakers www.e-t-a.com/us_home.html •
Excelsys Technologies www.excelsys.com • •
Fairchild Semiconductor www.fairchildsemi.com •
Falcon Electric, Inc. www.falconups.com •
Fuji Electric Corp. of America www.americas.fujielectric.com •
Gaia Converter www.gaia-converter.com • • •
General Electric www.geindustrial.com •
GlobTek, Inc. www.globtek.com • • •
Hammond Manufacturing Co., Ltd. www.hammondmfg.com •
Hioki USA www.hiokiusa.com •
HolyStone International — Holy Stone Enterprise Co., Ltd. www.holystonecaps.com •
Honeywell Power Products www.honeywellpower.com • •
Illinois Capacitor, Inc. www.illcap.com • •
IMS/AMCO Engineered Products www.amcoengineering.com •
Infineon Technologies Corp. www.infineon.com •
Integrated Power Designs www.ipdpower.com • •

AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT


Power Supplement Source Guide 15

DC/DC power supplies


AC/DC power supplies

Circuit protection

semiconductors
Cooling devices

Transformers
Capacitors
Batteries

Resistors
Discrete
devices
Company URL
International Manufacturing Services, Inc. www.ims-resistors.com •
Interpower Corp. www.interpower.com • •
Intersil Corp. www.intersil.com •
Isotek Corp. www.isotekcorp.com •
IXYS Corp. www.ixys.com •
IXYS Integrated Circuits Division www.ixysic.com •
Keithley Instruments, Inc. www.keithley.com •
Kepco, Inc. www.kepcopower.com • •
Keysight Technologies, Inc. www.keysight.com/find/power • • •
Kikusui America, Inc. www.kikusuiamerica.com •
KOA Speer Electronics, Inc. www.koaspeer.com • • •
Laird Technologies www.lairdtech.com •
Linear Technology Corp. www.linear.com • •
Littelfuse, Inc. www.littelfuse.com • •
Maxwell Technologies, Inc. www.maxwell.com •
Mean Well USA, Inc. www.meanwellusa.com • •
MEGA Electronics, Inc. www.megaelectronics.com • •
Memory Protection Devices, Inc. www.memoryprotectiondevices.com •
MicroPower Direct, LLC www.micropowerdirect.com • •
Microsemi Corp. www.microsemi.com • •
Mini-Systems, Inc. www.mini-systemsinc.com • •
Minmax Technology Co., Ltd. www.minmax.com.tw • •
Modular Devices, Inc. www.mdipower.com • •
Monolithic Power Systems www.monolithicpower.com •
Mornsun America, LLC www.mornsunamerica.com • •
MTM Power www.mtm-power.com •
Murata Electronics North America, Inc. www.murata.com • • • • • •
Murata Power Solutions, Inc. www.murata-ps.com • • •
NH Research, Inc. www.nhresearch.com •
NXP Semiconductors www.nxp.com/microcontrollers •
Ohio Semitronics, Inc. www.ohiosemitronics.com •
Ohmite Manufacturing Co. www.ohmite.com •
Okaya Electric America, Inc. www.okaya.com • •
ON Semiconductor www.onsemi.com •
Orion Fans — Knight Electronics, Inc. www.knightonline.com •
na.industrial.panasonic.com/products/
Panasonic Batteries
batteries •
P-DUKE Technology Co., Ltd. www.pduke.com • •
Pentair — Schroff http://www. schroff.pentair.com/en/schroff-na •
Phihong USA Corp. www.phihong.com •
Pico Electronics, Inc. www.picoelectronics.com • • •
Pioneer Magnetics www.pioneermagnetics.com •
Polytron Devices, Inc. www.polytrondevices.com • •
Power Mate Technology, Inc. www.powermateusa.com •
Powerbox USA, Inc. www.powerbox.se • •

ELECTRONIC PRODUCTS POWER SUPPLEMENT • electronicproducts.com • AUGUST 2017


16 Source Guide Power Supplement
Advertisers of this supplement are shown in red.

DC/DC power supplies


AC/DC power supplies

Circuit protection

semiconductors
Cooling devices

Transformers
Capacitors
Batteries

Resistors
Discrete
devices
Company URL
Powerex, Inc. www.pwrx.com • •
Prem Magnetics, Inc. www.premmagnetics.com •
Premier Magnetics, Inc. www.premiermag.com •
Protek Power North America www.protekpowerna.com •
Qualtek Electronics Corp. www.qualtekusa.com • • •
RECOM Power www.recom-power.com • •
Renco Electronics, Inc. www.rencousa.com •
Renesas Electronics America, Inc. www.am.renesas.com •
ROHM Semiconductor www.rohm.com • • • • • •
Rosenberg USA, Inc. www.rosenbergusa.com •
Saft America, Inc. www.saftbatteries.com •
SCHURTER, Inc. www.schurterinc.com • •
Semiconductor Circuits, Inc. www.dcdc.com • • •
SL Power Electronics Corp. www.slpower.com • •
Stanford Research Systems, Inc. www.thinksrs.com •
State of the Art, Inc. www.resistor.com •
STMicroelectronics, Inc. www.st.com •
Sumida America Components, Inc. www.sumida.com • •
Syfer Technology Ltd. — A Dover company www.syfer.com •
Tadiran Electronic Industries, Inc. www.tadiranbat.com •
Taiyo Yuden (U.S.A.), Inc. www.t-yuden.com •
TDK Corporation of America www.tdk.com • • •
TDK-Lambda Americas, Inc. — AC-DC and DC-DC Power Supplies www.us.tdk-lambda.com/lp • •
TDK-Lambda Americas, Inc. — Programmable and High-Voltage
Power Supplies
www.lambda-hp.com •
Tectrol, Inc. www.tectrol.com • •
Texas Instruments, Inc. www.ti.com • • •
The Bergquist Co. www.bergquistcompany.com •
Thermacore, Inc. www.thermacore.com •
Toshiba America Electronic Components, Inc. www.toshiba.com/taec •
Total Power International, Inc. www.total-power.com • •
TRACOPOWER www.tracopower.com • •
Triad Magnetics www.TriadMagnetics.com •
Tri-Mag, LLC www.tri-mag.com • • • •
Tumbler Technologies + TRUMPower www.trumpower.com • •
UltraVolt, Inc. www.ultravolt.com •
Unipower Corp. www.unipower-corp.com • •
United Chemi-Con, Inc. www.chemi-con.com •
VARTA Microbattery, Inc. www.varta-microbattery.com •
Vicor Corp. www.vicorpower.com • •
Vishay Intertechnology www.vishay.com • • • • •
Wall Industries, Inc. www.wallindustries.com • •
Wolfspeed www.wolfspeed.com •
XP EMCO www.emcohighvoltage.com • •
XP Power www.xppower.com • •
Zettler Magnetics, Inc. — Zettler Comps. www.zettlermagnetics.com •
AUGUST 2017 • electronicproducts.com • ELECTRONIC PRODUCTS POWER SUPPLEMENT
CHARGE UP
YOUR SUMMER

b a t t e r y h o l d e r s . c o m

MPD
m e m o ry p r o t e c t i o n d e v i c e s
A57_ElecProds7_875x10_5.qxp_Layout 1 6/28/17 11:39 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 almost 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


Think Pico Small - Surface Mount & Thru Hole
Over 5000 std Ultra Miniature
• Ultra Miniature Designs
Military • COTS • Industrial
• 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: [email protected]  •  FAX: 914-738-8225

VISIT OUR EXCITING NEW WEBSITE: www.picoelectronics.com

You might also like