www.newelectronics.co.

uk
www.newelectronics.co.uk

1128July
June2017
2016

SPACE
SMART
ROVERS ELECTRONICS
TRANSPORT
EXCEED BATTERY
EXPECTATIONS SAFETY &MANAGEMENT
SECURITY
REDEFINING ENCLOSURES
HARDWARE IN REAL
EDA & MACHINE
TIME EMC SHIELDING
CATCHING LEARNING
UP WITH CLOCKS

THE INTERFACE
OF CHOICE?
The emergence of smart assistants
and speakers is driving major changes
in the audio market

THE REAL DEAL?

Are augmented and virtual reality
nally set to become the next big
computing platforms?

001_NELE_JUL11.indd 1 03/07/2017 12:04

002_NELE_JUL11 03/07/2017 10:41 Page 1
 New Electronics weekly eZine features the
latest blogs, news, articles and more.
To register for your copy, go to
www.newelectronics.co.uk CONTENTS VOL 50 NO 13

18 27

14 21 30

COMMENT 5 INTERVIEW 12 SYSTEM TEST 24

The UK needs to look beyond
graphene and chart a route
to the mass production and
application of 2D materials
NEWS
Graphene based SpinFET
could integrate data
processing and storage
operations 7 500 smart devices per house in the near future.
Knowledge Transfer Network
establishes SIG to promote
commercial opportunities for
2D materials 8 Harnessing the power of big data
The Government launches a
100million fund designed
to attract highly skilled
researchers 9
The tricks being used to
extend the life of optical
lithography are piling up the
chip design challenges 10
A passion for training
A lack of proper RF training encouraged Oren
Hagal to set up his own company in 2000 to
help engineers prepare for the design challenges
they face
COVER STORY
The interface of choice
Market researchers expect there to be more than
Controlling them using conventional interfaces will be
challenging, so will voice become the new click?
SMART CITIES
While the case for the smart city is well known,
the infrastructure is still evolving and real
innovation will be needed if smart cities are to
become reality
SMART TRANSPORT
The long haul
Sophisticated truck automation and connectivity
solutions exist, but legislation is needed before
we see autonomous commercial vehicles being
used on the road.
Microbumps: the probing challenge
In order to attain high yields, manufacturers of
3D stacked ICs must test dice before packaging
them. Its a challenging problem, but a solution
has been developed
14 POWER 27
The power to manage
Battery management systems continue to
support the drive towards higher reliability in
electric and hybrid vehicles and wireless
communications are set to help
18 EDA 30
Learning by example
Using machine learning techniques to automate
some repetitive chip design processes could play
a leading role in getting bug free parts to market
more quickly
21
MISSION STATEMENT
New Electronics keeps designers and
managers abreast of the latest developments
in the worlds fastest moving industry

www.newelectronics.co.uk 11 July 2017 3

Realise your vision
We have all the parts you need to solve your technical
challenges and bring your design projects to life.

Never get stuck again. Our extensive range

of semiconductors, passives and interconnects
has got you covered. With thousands Find the right parts at
of brands you know and trust, you can
count on us for the best solution. uk.rs-online.com

4 11 July 2017 www.newelectronics.co.uk

 COMMENT 2D MATERIALS

Beyond
graphene
MAPPING A ROUTE TO MASS
PRODUCTION OF 2D MATERIALS

S ince the discovery of graphene in 2004, the fascination around the

so called wonder material has grown to include a whole range of 2D
materials.
To date, 500 2D materials have been developed by teams around the
world and the market for these materials is thought to be worth 300million
a year; a figure that is expected to rise as more devices and products feature
these materials.
Robert Quarshie, head of materials at the Knowledge Transfer Network
(KTN) and author of the report Creating Value from Non-carbon 2D Materials
Beyond Graphene, argues that the UK is well positioned to take a
significant share of this market, but needs to maximise its opportunities.
Barriers to commercialisation are considerable, however, including a lack
of skills and relevant funding, poor engagement with potential end users and
customers and a misaligned supply chain landscape.
While the likes of Samsung and Intel dominate the international patent
landscape, SMEs and universities are doing the work in the UK, with a focus
on research and materials manufacture.
While there has been considerable focus on applications, there is a real
need to improve manufacturing and the report says more work needs to
be done in the UK to scale up manufacturing and promote the use of 2D
materials, as well as to improve manufacturing.
So what does the UK need to do to take advantage of its position? The
report calls for: the creation of an Industry Challenge to accelerate the
development of the supply chain; funding over an extended period for centres
of excellence to scale up production; and support for scaled demonstrators to
show what graphene is capable of.
A range of industries have expressed an interest in and recognised the
benefits of these new materials and, now that work has been done to address
graphenes lack of band gap the property that makes silicon so useful in
digital electronics it is an attractive choice for digital circuits.
The KTN has launched a special interest group (SIG) to focus on
commercial opportunities, bringing together manufacturers of 2D materials
and those looking to develop applications.
Opportunities abound, but the SIGs work will be crucial if challenges are
to be overcome and a clear route to mass production developed.

Neil Tyler, Editor (neil.tyler@markallengroup.com)

www.newelectronics.co.uk 11 July 2017

 NEWS GRAPHENE SPINFET

A new spin on graphene

Editor Neil Tyler GRAPHENE BASED SPIN FET COULD INTEGRATE DATA PROCESSING AND
neil.tyler@markallengroup.com
STORAGE OPERATIONS. GRAHAM PITCHER REPORTS.
Online Editorial Peggy Lee
Assistant peggy.lee@markallengroup.com Graphene Flagship researchers based at Chalmers University of Technology have produced working
Consulting Graham Pitcher versions of a graphene-based spin field-effect transistor operating at room temperature and say this
Editor graham.pitcher@markallengroup.com is a step towards integrating spintronic logic and memory devices.
Contributing Chris Edwards Graphene is an excellent medium for spin transport at room temperature, due to its low atomic
Editors John Walko mass, said Saroj Dash, group leader and an associate professor at Chalmers. However, an
editor@newelectronics.co.uk unsolved challenge was to control the spin current at ambient temperature.
Art Editor Andrew Ganassin Graphene has been shown to transport spin over long distances by several Flagship Groups. By
andrew.ganassin@markallengroup.com combining graphene with another layered material in which spin doesnt last as long, it is possible
Illustrator Phil Holmes
to produce a device which operates like a spin field-effect transistor.
Dash added: By combining graphene, where spin lasts for nano seconds, with molybdenum
Sales Manager James Creber
disulphide, where spin only lasts for picoseconds, you can control where the spin can go by using
james.creber@markallengroup.com
a gate voltage essentially you can create a spin switch. Importantly, we show in this research a
Publisher Peter Ring
peter.ring@markallengroup.com
particular materials mix which enables this spin-switch to work at room temperature.
While many layered materials are said to be promising for spintronics, the team says the bigger
Managing Jon Benson goal is to create novel spin phenomena by stacking different layers with complementary properties.
Director jon.benson@markallengroup.com

Production Nicki McKenna

Manager nicki.mckenna@markallengroup.com

New Electronics editorial advisory panel

Trevor Cross, chief technology ofcer, Teledyne e2v
Pete Leonard, electronics design manager, Renishaw
Pete Lomas, director of engineering, Norcott Technologies
Neil Riddiford, principal electronics engineer, Cambridge
Consultants
Adam Taylor, embedded systems consultant
ISSN 0047-9624 Online ISSN 2049-2316
Annual subscription (22 issues):
UK 108. Overseas; 163. Airmail; 199.
New Electronics, incorporating Electronic Equipment News
and Electronics News, is published twice monthly by
MA Business, Hawley Mill, Hawley Road, Dartford, DA2 7TJ.
T: 01322 221144 E: ne@markallengroup.com
Moving on?
Raspberry Pi wins MacRobert Medal
The team behind Raspberry Pi has won the Royal Academy of
Engineerings MacRobert Medal, one of the most prestigious
engineering awards made by a UK organisation. Along with
the gold medal, the team also won 50,000.
Pete Lomas, director of engineering with Norcott
Technologies and designer of the Raspberry Pi, said: Im so
pleased for the team and every member of our community that
made it possible.
More than 14million Raspberry Pis have been sold since
the device was launched in February 2012, making it the most
successful UK developed computer. Originally intended to
capture the attention of school students, more than half of the

Pic: Raspbe
If you change jobs or your company moves, please contact Raspberry Pis sold have been used in industrial applications.
circulation@markallengroup to continue receiving your
free copy of New Electronics

Voice interface deal signed

2017. All rights reserved. No part of New Electronics may
be reproduced or transmitted in any form, by any means,
electronic or mechanical, including photocopying, recording or
any information storage or retrieval system, without permission
in writing from the Publisher.The views expressed do not
necessarily represent those of the editor of New Electronics.
Advertisements in the journal do not imply endorsement of the
products or services advertised.
Origination: CCM, London. Printed by Pensord.
XMOS has acquired US based Setem Technologies, the developer of Advanced Blind Source Signal
Separation technology. This acquisition allows us to extend the development and deployment of
world-leading algorithms capable of driving the rapid adoption of voice user interfaces, said Mark
Lippett, XMOS president and CEO.
One of the problems with voice interfaces is that sources can be 5m from the microphone or as
close as 1m and both are subject to competing conversations, reverberation and background noise
the so called cocktail party problem.
Traditional beamforming technologies can struggle to identify specific sound sources in noisy
environments. Setems solution uses a combination of signal processing and machine learning
techniques based on voice biometrics. By dissecting and reconstructing the sound field in the time,
frequency and spatial domains and in real time, Setem enabled microphones can extract individual
voices from background noise.
We are excited about joining forces with XMOS, said Jochen Meissner, Setems CEO, who will
become general manager of XMOS Boston office and an XMOS director. We have been working
with XMOS for more than 18 months and have concluded that the xCORE architecture is the perfect
match for our patented algorithms.
For more on voice interfaces, see the Cover Story on p14.

www.newelectronics.co.uk 11 July 2017 7

 NEWS
Narrowest
bandwidth yet
Researchers from the University of Twente
have developed a diode laser on a chip with a
bandwidth of 290Hz. According to the team,
this is the most accurate such device yet
created.
The hybrid device with two photonic
chips connected optically is tuneable, which
means that users can select the particular
wavelength, within a broad range.
The researchers believe their development
will bring countless applications within reach,
including the control of movable antennae on
phone masts for 5G mobile internet, faster
data flows through fibre networks, more
accurate GPS systems and better sensors for
monitoring structural integrity.
SQA for next gen
comms test
Looking to offer higher levels of integration,
Anritsu has released the MP1900A, a signal
quality analyser supporting next generation
telecoms, data centre and storage network
transmission standards, including PCIe Gen 4,
Thunderbolt Gen 3 and USB Gen 3.1.
As of today, no other test and
measurement company has a platform that
tests all these new technologies at the same
time, said Alessandro Messina, EMEA
wireline business marketing senior manager.
The new measurement platform features
16 testing channels and eight slots. Said
to be an all-in-one solution, the device can
measure high speed interface designs during
the early development stage, including
optical transceivers used in servers and
communications equipment, and optical
transceivers/modules for M2M and IoT
applications.
Our platform is already ready for PCIe
generation 5, which will come in one or two
years from now, Messina added.
8 11 July 2017 www.newelectronics.co.uk
2D MATERIALS
SIG to promote 2D materials
KNOWLEDGE TRANSFER NETWORK TO PROMOTE COMMERCIAL OPPORTUNITIES FOR
GRAPHENE. NEIL TYLER REPORTS
The Knowledge Transfer Network (KTN) has launched a special interest group to promote
commercial opportunities for graphene and other 2D materials, which have a market thought to
be worth in excess of 300million a year.
Robert Quarshie, head of materials at the KTN and author of the report Creating Value
from Non-carbon 2D Materials Beyond Graphene said the UK is well positioned to take a
significant share of this market but needs to maximise its opportunities.
The report warns that barriers to commercialisation are considerable, including a lack of
skills and relevant funding, poor engagement with potential end users and customers and a
misaligned supply chain. But it contends the UK is in a position to maximise opportunities in
applications as diverse as wearable technology, energy storage and medical devices.
The UK industry needs to be open about their trends and drivers and as such need to work
a lot more closely with researchers in developing technology roadmaps, Quarshie suggests.
Our report calls on the creation of an Industry Challenge to accelerate the development of the
supply chain; funding over an extended period for centres of excellence to scale up production
and support for scaled demonstrators to show just what graphene is capable of doing.
Ferroelectric layer for
precise control
Devices created using the conducting polymer
PEDOT:PSS can switch rapidly at a threshold
voltage, say researchers from Linkping
University. This could enable control of the
current state of printed electronic transistor
circuits and displays, as well as the colour state of individual pixels within displays.
However, the lack of any threshold in the redox-switching characteristics of PEDOT:PSS
hampers bistability and rectification; characteristics that would allow for passive matrix
addressing in display or memory functionality said senior lecturer Simone Fabiano.
To resolve this issue, the team applied a thin layer of a ferroelectric material onto one
electrode in organic electrochemical devices and circuits.
The thickness of the layer determines the voltage at which the circuit switches or the display
changes colour. Transistors are no longer required in the displays: we can control them pixel-
by-pixel simply through a thin ferroelectric layer on the electrode, Fabiano explained.
York EMC acquired IoT development platform
In a move which is said to underpin its ability Cypress and Arrow have launched a
to provide regulatory compliance services, development platform designed to help
York EMC Services (YES) has been acquired engineers get connected IoT products to
by Eurofins Product Testing. market more quickly.
Nick Wainwright, YES CEO, noted: YES The rst Quicksilver kit featuring
has grown rapidly over the last five years Cypress WICED platform and its CYW43907
and we have made significant investments 802.11n MCU includes the WICED Studio
in our laboratories and our products and SDK and Arduino-compatible headers for
services. By joining Eurofins, we can ensure expansion capability. Also provided are
this momentum is maintained and that our temperature, humidity and three axis motion
business continues to develop. sensors, enabling the design of complete IoT
Eurofins Group provides analytical services edge devices for a range of markets.
through an international network with more The companies say they will launch a
than 300 laboratories in 39 countries and second Quicksilver kit, supporting 802.11ac
28,000 staff. Wi-Fi, in Q4 2017.
 NEWS UK RESEARCH

Fund targets Leti, Fraunhofer

partnership
best researchers Leti and Fraunhofer are collaborating to
develop next generation microelectronics
technologies and believe their partnership
THE GOVERNMENT PUTS 100MILLION INTO ATTRACTING HIGHLY SKILLED will spur innovation in their countries, while
RESEARCHERS TO THE UK. NEIL TYLER REPORTS. strengthening European strategic and
economic sovereignty.
A 100million Global Talent Research Fund to attract highly skilled researchers to the UK has Letis CEO Marie Semaria said: Building
been named after the Nobel Laureate Ernest Rutherford. The fund, which will provide fellowships on previous successful collaborations, Leti
for early career and senior researchers, will be administered by Innovate UK and the Research and the Fraunhofer Group for Microelectronics
Councils until UK Research and Investment (UKRI) is formed in 2018. will bring complementary strengths to the
Looking to attract researchers from the developed world and from emerging research task of keeping France and Germanys
powerhouses such as India, China, Brazil and Mexico, ministers hope the fund will help to microelectronics industries in the forefront
maintain the UKs position as a world-leader in science and research. and offer innovations across Europe.
Speaking at the funds launch, Universities and Science Minister Jo Johnson said: Research Collaborative R&D projects will include:
and innovation is at the heart of the governments Industrial Strategy. According to Johnson silicon-based technologies for next-generation
the Prime Minister is determined to make the UK attractive for scientists, innovators and tech CMOS processes and products; extended
investors. She wants us to be a country that attracts the brightest and best minds, he said. More than Moore technologies for sensing and
Johnson added the Rutherford Fund is intended to send a strong signal that, even as the communication applications; and advanced
UK leaves the EU, it will remain open to the world and reinforce the Governments ambition of packaging technologies.
making the UK the go-to country for innovation and discovery.

Performance
In the next issue of New Electronics, the recently appointed Chief Executive of UKRI Sir
Mark Walport outlines how UKRI will support the UKs Industrial Strategy.

Superstretchable
supercap
A stretchable and compressible
supercapacitor developed by researchers
at the City University of Hong Kong can
be stretched to 1000% in length and
compressed to 50% in thickness without
breakage, cracking or other damage.
The researchers developed a
polyelectrolyte composed of a polyacrylamide
hydrogel reinforced with vinyl-functionalised
silica nanoparticles. This material is made
stretchable by cross-links in the vinyl-silica
nanoparticle and conductive because the
polyelectrolyte, holds and transfers ions.
LiS battery test centre
OXIS Energy has set up an International
Lithium Sulphur battery system test centre.
Said to be the first of its kind in Europe,
the facility will enable OXIS to further its
development of LiS chemistry and technology.
The test centre, which will enable OXIS
to develop new processes for the use of
rechargeable LiS battery systems, will also
be available to OXIS customers. Meanwhile,
the company will continue to work with its test
partner the University of South Wales.
The centre will house specialist equipment
to test cells and battery systems in order to
meet international standards and specific
industrial sectors regulatory requirements.
and value
According to Rigol Technologies, its Phoenix
oscilloscope chipset will allow the company
to bring its value proposition to more
performance oriented applications.
The Phoenix chipset has three elements.
The analogue front end supports bandwidths
of up to 4GHz and integrates the functionality
required by a digital oscilloscopes analogue
modules. A signal processing chip supports
the acquisition of data at rates of up to
10Gsample/s, while the probe amplier chip
supports differential probes with bandwidths
of up to 6GHz.
Amongst the benets predicted
for devices featuring the chipset are
signicantly faster waveform capture rates,
new ltering and triggering capabilities and

Smart bandage to monitor wounds

unprecedented memory depths and search
capabilities.

A high-tech system that supplies nursing staff with relevant data about the condition of a
wound has been developed by researchers from Empa, ETH Zurich, the Swiss Center for
Electronics and Microtechnology and the University Hospital Zurich.
With chronic wounds, nursing staff have to change the dressing regularly not just for
reasons of hygiene, but also to examine the wound. The researchers believe it would be
better to leave the bandage on for longer and have the nursing staff read the condition of
the wound from outside.
A bandage designed as part of the Flusitex project features sensors incorporated into
the base material. These glow with different intensity if the wounds pH level changes. And
if certain substances appear in the wound uid, uorescent sensor molecules respond by
glowing. Using a colour scale, the quantity of the substance be deduced.

www.newelectronics.co.uk 11 July 2017 9

 NEWS ANALYSIS 7NM MANUFACTURING
New node,
new challenges
AS COMPANIES PLAY TECHNOLOGY LEAP FROG, 7NM
PUSHES DESIGN TO NEW LIMITS. CHRIS EDWARDS REPORTS.
While chipmakers are keen to move to the 7nm node, the use of
tricks to extend the life of optical lithography, amidst a continuing wait
for EUV, is piling up the design challenges.
In the last five to six years, we have gone through six to seven
nodes, starting with variants of 28nm, said Balaji Velikandanathan,
quality engineer at Qualcomm, in a panel session at Junes Design
Automation Conference. We are talking about [moving from] inception
to tapeout in nine months and [introducing] a new process node every
year.
Qualcomm wants to move to 7nm to push frequency to 3GHz and
reduce power by 30%, compared to the 10nm Snapdragon 835 used in
phones such as the HTC U11.
But the need to use self-aligned double patterning (SADP) to form
the densest metal interconnect layers brings with it complex new
design rules, as well as much higher resistance in the wiring itself.
With SADP, metal lines have to run parallel to each other on each layer.
Any connection that includes one or more turns needs to use at least
two metal layers, incurring the resistance cost of vias, as well as the
very thin metal wires.
To work around high resistance in signal lines that need high
switching speeds, foundries have adopted via pillars or via ladders and
encourage the use of the coarser metal on upper layers for high-speed
signals that would be prone to electromigration if routed on the finer
wiring layers. The pillars use two vias in parallel to move up each layer,
which reduces overall resistance.
The via pillar doesnt come for free, said Velikandanathan. It
uses additional routing resources and places additional challenges on
the router.
Foundries such as GlobalFoundries (GF) and TSMC have involved
designers in the creation of their 7nm processes in order to work
out how to make designs economic and to claw back drops in silicon
utilisation incurred by structures such as via pillars. There is no free
10
lunch, said GFs design-enablement
fellow Greg Northrop. You have to make compromises but, ultimately,
everybodys real interest is making sure the final cost is minimised.
A major impact is that we end up with a significantly scrunched
standard cell and this will impact the way in which libraries interact
with the router.
One of the biggest problems for design with standard cells is
providing access points for the router to the pins of each gate. This
problem is exacerbated by rules governing SADP that make small jogs
in local routing extremely expensive. Synopsys, for example, found
in early work with GF that it often makes more sense to move cells
around so an interconnect can use a single straight line.
A future shift to EUV patterning for the lowest metal layers may
relax the rules for designers, making it easier to hit density and speed
targets. Samsung said at the VLSI Technology Symposium that use of
EUV in critical layers reduced routing congestion and improved speed
by removing the excess parasitic capacitance of the dummy metal lines
needed by SADP-based processes.
Willy Chen, deputy director of TSMCs design and technology
platform, said its N7+ process which will use some EUV-imaged
layers should reduce area consumption by 10%. However,
Northrop said while GF sees EUV as a potential vehicle for reducing
manufacturing cost, it wants to minimise disruption to flows and is
unlikely to alter the design rules.
One issue that foundries face is that EUV, although potentially a
better lithographic technology, changes the process flows in those
layers dramatically.
From a lithography perspective, EUV is wonderful, said David
Fried, CTO of Coventor. From a process-integration perspective, it has
many implications.
11 July 2017 www.newelectronics.co.uk
 INTERVIEW OREN HAGAI
A passion
for training
A lack of proper RF training encouraged Oren
Hagai to set up his own company in 2000 to help
prepare engineers for the design challenges they
face. By Neil Tyler.
A
recognised expert in RF and Microwave Oren Hagai, president
& CTO of Interlligent RF & Microwave solutions, is passionate
about training and providing the right learning experience for
the upcoming generation of RF engineers.
The move towards 5G has raised a lot of questions around training
and the importance of preparing engineers when it comes to the
challenges associated with designing applications using millimetre-wave
frequencies, he says.
After I graduated, I worked for a number of start-ups and it wasnt
long before I realised there was a lack of appropriate training in the
RF community. The focus was on encouraging new engineers to study
and read about the subject, but there was no real methodology or
curriculum to deliver the skill sets or knowledge to work effectively in
this space. As a design engineer I didnt feel I was being prepared for
real life, so I decided to set up Interlligent.
From the beginning, the aim was to provide practical in-depth
training that introduced engineers to sound engineering practices.
Initially, the company provided training only, but it has developed to
address a range of complementary fields.
When I and a couple of colleagues set up the company, we saw it
simply as a way of funding our academic studies. I was studying for a
Masters at the time and it wasnt until a few years later when we won
a significant contract that not only involved training but also required us
to acquire test equipment as part of the deal that we began to realise
we had a real business on our hands, he says.
We needed to supply test equipment to provide engineers with
hands-on experience and, after the telecoms crash, there was plenty
of surplus equipment on the market. That lead us to renting out test
equipment.
Fast forward to 2017 and Interlligent now employs 50 people
and is recording year on year growth in excess of 15%. The company
comprises five business divisions (including engineering services and
company representation) and operates in the UK, having registered as a
UK company in 2014.
Not all of our divisions are active in the UK, but with so many
12
interesting UK companies, I think it is important that we increase our
presence there.
When it comes to the test and measurement rental market, Hagai
sees the IoT as well as the need to move towards millimetre-wave
frequencies for developing prototype 5G components and subsystems
as creating a host of opportunities and challenges.
Start-ups and smaller companies particularly like to use rental to
ease their budgets, he suggests.
Additional services, such as training and loan of accessories, are
also an advantage to them when renting, rather than buying outright.
While the rental market has been the main business driver, training
has helped to establish the company as a centre of excellence.
Our approach has always been hands on, providing in-depth
training for all levels of experience when it comes to RF. We have a
large facility in Tel Aviv which not only supports local businesses, but
also attracts engineers from around the world.
Were training hundreds of engineers every year and that has
helped us establish a level of trust with companies that many of our
competitors find hard to emulate, Hagai believes. We also offer
companies training in-house and often fly engineers from Israel to
deliver our courses.
In 2016, Hagai changed his role from CEO to CTO in order to spend
more time developing in depth training materials and to teach advanced
microwave subjects.
The advent of 5G has transformed the market and modern
designers and engineers are looking to acquire a more detailed
knowledge of waveguides, he contends. Demand for training
in mmWave technologies, software defined radio and data
communications has exploded and, when it comes to mmWave, there
are techniques and tricks associated with managing distortion that
conventional engineers need to better understand.
There are plenty of skilled engineers. They just lack the necessary
knowledge about the latest mmWave technologies. Everything in this
space is extreme and our role is to accelerate awareness and learning.
While universities are providing the theory, our aim remains, as it
always has been, to provide a practical approach and train engineers
for the industry.
Our focus on training has helped to differentiate our brand from
our competitors, emphasising our technical knowledge and expertise.
Engineers, especially those we work with in the UK, appreciate that.
Interlligent has also established an RF incubator in Israel and is
interested in developing a similar service in the UK. There is certainly
potential in the UK and Id like to replicate what we have done in
Israel, Hagai suggests.
The Tel Aviv incubator offers entrepreneurs and RF related start ups
the opportunity to hire a facility fully equipped with test equipment. It
helps them to cut their costs significantly, but not only that, it enables
us to invest in suitable businesses and we have a portfolio of eight
companies in which we have taken a stake our involvement is not
only financial, but also managerial.
As a business, we are managed and run by microwave engineers
and that certainly sets us apart, Hagai concludes.
11 July 2017 www.newelectronics.co.uk
 Oren Hagai
Oren Hagai founded Interlligent in 2000.
After 16 years as CEO, he changed his role
to CTO in order to devote more time in
developing in depth training materials and
prepare the company for building its own
core technologies. An expert in the eld
of RF and Microwave engineering Hagai
studied electrical engineering at Tel-Aviv
university and served as a Microwave
Expert in Israels Defence Force.

www.newelectronics.co.uk 11 July 2017 13

I The interface
n the last two years, consumers
have become increasingly discerning
when it comes to audio quality

of choice
and expect to enjoy full compatibility
between their devices, regardless of
the brand.
A report commissioned by
Qualcomm The State of Play
found consumers were willing to
pay for products that met those
requirements and that, in an already
crowded market, manufacturers and
developers who looked to address
those needs would benefit.
Improving sound quality and
compatibility are providing a vital
competitive edge, but while audio
technology has come a long way in
a relatively short time, consumer
expectations continue to grow apace.
Qualcomm Technologies recently
used a developers conference in
China to unveil what it described Major changes are happening in the audio
as its next generation direct digital
market, but a key driver has been the emergence
feedback amplifier (DDFA) audio
technology. of smart assistants and speakers. By Neil Tyler.
Targeted at high-resolution audio
devices, such as wireless speakers,
soundbars, networked audio and
headphone amplifiers, DDFA is means many more OEMs will be the prospect of changing the way in
intended to replace more traditional able to integrate high-performance which we interact with our electronic
class D amplifiers with an all-digital amplifier capability into their products. Below: The devices.
pulse width modulator, while its The audio market is being AcuEdge Voice has traditionally been seen
closed-loop architecture means that transformed, whether thats how Development Kit as a very complex technology and
it can compensate for nonlinearities consumers access their music to high from Microsemi only available as a solution to those
of power supplies and output stages. resolution audio. can deliver companies with deep pockets and the
The result is that it can deliver much Were also seeing a dramatic enhanced audio super expertise to implement it.
higher fidelity audio and greater increase in the demand for voice processing Today, its a technology that is
design flexibility, while retaining the technology on the back of the capabilities more widely available. Word error
advantages traditionally associated emergence of smart assistants and rates have fallen significantly
with class D amplifiers. speakers, such as the Home and and machines are
Anthony Murray, general manager, Echo, he says. quickly becoming as
voice and music, with Qualcomm accurate as humans.
Technologies, said the ability to Voice recognition As a result,
deliver an improved sound experience Demand for voice voice is expected to
was a major focus for audio recognition encourage a host of
technologists. technology is new products, as well
DDFA is intended to help growing and, as improve existing
customers who are having to meet while it first devices, according to
Image: Adobestock

the needs of increasingly audio-savvy appeared Murray.

consumers, he suggested. in analogue Sales of voice-activated
While DDFA has previously form back in digital assistants, such as Google
been used by premium audio the 1950s, its a Home or the market leading Amazon
manufacturers, this new release technology that holds Echo, have been running at around

14 11 July 2017 www.newelectronics.co.uk


5million units per year, and market
analysts are forecasting annual
sales of more than 10m in 2017.
The overall smart speaker/digital
assistance market could reach 47m
units by 2022.
While one significant company,
Apple, has been accused of being late
to the market, it has recently started
work on a Siri-based device. Reports
suggest it will be a voice-controlled
speaker capable of responding to
basic commands and queries but
which, according to Bloomberg News,
will have one key differentiator from
those products already on the market
audio quality.
It is suggested that the speaker
will offer virtual surround sound
technology, using one speaker to
create the impression that the sound
is coming from several devices.
While Amazon and Google lead the
market, Microsoft, IBM and Nuance
are developing voice controlled
products whose quality has improved
significantly, allowing users to talk
more naturally. Meanwhile, many
companies are looking to enable
complete residential IoT solutions.
Automatic Speech Recognition
(ASR) is defined as the ability of a
machine or program to receive and
interpret dictation or to understand
and carry out spoken commands,
explains Shahin Sadeghi,
Microsemis director of marketing and
applications. ASR Assist uses audio
enhancement to target the human-to-
machine voice interface specifically.
Speech or voice recognition is
a computer software program or
hardware device with the ability to
decode the human voice and uses
ASR software.
Most of these programs require
the device to be trained to recognise
a voice so that it can convert that
speech into data accurately, Sadeghi
says.
A variety of voice recognition
systems, including speaker
dependent systems, require a
series of words and phrases to be
www.newelectronics.co.uk 11 July 2017
pre-recorded. But there are also:
speaker independent systems, in
which no training is required; discrete
speech recognition systems, in
which the users may have to pause
between words to enable the speech
recognition software to identify each
separate word; and continuous
speech recognition, where the system
can understand a normal rate of
speaking. Its this latter space where
there has been real improvement in
recent years.
The performance of voice
controlled products is markedly
better, says Sadeghi. Key drivers
have been set top boxes and smart
TVs, where voice commands can be
complex.
But whether its Amazon, Google
or Apple, all the big players are now
enabling more complete ecosystems.
The rise of smart speakers has
accelerated over the past few years
and they are now seen, according to
Sadeghi, as hubs for smart homes.
But voice recognition and voice
direction is not only targeting the
consumer market.
Increasingly, voice is being used
across industries ranging from
healthcare and telecommunications to
the military and personal computing.
It is being used in a growing number
of commercial settings to improve the
performance of employees and being
used to transform data into speech to
instruct workers.
COVER STORY VOICE INTERFACE
TECHNOLOGY
Voice is no longer seen as
research or as an interesting science
project; demand for voice recognition
technology is surging and opening up
a host of opportunities for smaller
businesses and a growing number of
companies are releasing development
boards, processors and low power
microphones to meet that demand.
In June, Microsemi unveiled its
AcuEdge Development Kit for the
Amazon Alexa Voice Service (AVS).
The performance AcuEdge can deliver enhanced
of voice controlled audio processing and is intended
products is to improve voice recognition
markedly better. rates in adverse or difficult audio
Key drivers have environments, Sadeghi explains.
been set top We are seeing demand for voice
boxes and smart recognition accelerating and are
TVs, where voice looking to provide a much faster
commands can and easier way to prototype devices
be complex. and to assist in the development of
Shahin Sadeghi human to machine (H2M) applications
being driven by the IoT and automated
assistance markets.
The kit will enable third party
developers and manufacturers to
evaluate and incorporate Alexa
functionality into any H2M application
and does this by interfacing with
Below: Companies Microsemis Timberwolf ZL38063
like Microsemi, multi-mic audio processor.
Qualcomm and Qualcomm Technologies has
XMOS want to help also unveiled what it calls its Smart
manufacturers Audio Platform. We want to help
accerate the manufacturers accelerate the
commercialisation of development and commercialisation
smart speakers of smart and networked speakers,
Murray explained.
The platform is flexible. It offers
two SoC options based on our
APQ8009 and APQ8017 devices
and, when combined with a range of
software configurations, means that
OEMs will be able to create smart
speaker systems for multiple products
and categories.
Its an integrated platform that
brings together processing capability,
a variety of connectivity options, voice
user interfaces and high end audio
technologies. We are finding that
users are demanding highly intuitive
smart speakers.
15
 COVER STORY
A world of gadgets
Voice-activated gadgets can range
from those with basic sound-
control capabilities to smarter ones
controlled by a combination of voice
and gestures. Children-friendly
gadgets are also finding their way on
to the market, says Murray.
The market is broad and includes
devices as varied as: smart helmets,
using voice to control display
systems; remote controls and alarm
clocks that can be controlled via
various voice commands; voice
activated smart phones and lights;
and the electronic assistants and
applications mentioned earlier.
Voice-control is by far the
simplest means of guiding technology
and we are seeing more devices
becoming available to this type of
technology, according to Murray.
Qualcomms integrated voice
solution has been designed to deliver
advanced multi-mic far-field voice
capabilities that can combine highly
responsive voice activation with
beamforming technologies.
The platforms voice software
incorporates echo-cancellation,
noise suppression and barge-in
capabilities, supporting a more
reliable voice interface in loud or noisy
environments even when users are
a distance from the smart speaker,
says Murray.
As with most electronic devices, a
key challenge remains power. When
it comes to speech recognition, one
thing that designers are confronting
and which, for many, is a radically
new challenge is that their designs
need to support active all day
speakers, says Murray.
That is a real technical challenge
and battery life remains a serious
issue. While playback times are
now running at eight hours or
more double what they were a
few years ago the trade-offs have
been considerable. We are currently
looking at not only new battery
chemistries, but also the actual
shape of batteries. While I cant say
16
VOICE INTERFACE TECHNOLOGY
more at this stage, new products
in the pipeline will reduce power
requirements significantly.
Vesper, a sensor specialist, has
developed what it believes is the first
MEMS microphone that can bring
voice activation to battery-powered
consumer devices, and does so by
drawing nearly zero power.
The companys ZeroPower
Listening MEMS microphone, the
VM1010, is a tiny piezoelectric
MEMS microphone that allows
product designers to offer touchless
user interfaces to consumers and
consumes just 6A while in listening
mode, waiting for a keyword.
Power issues
Consumers want to interact with
their electronic devices using
just their voice and there is an
explosion in voice-enabled consumer
electronics, including smart speakers,
smart earbuds, and diverse IoT
products, said Matt Crowley, the
companys CEO. The problem is that
speech consumes a lot of power,
1000W or more, and that is why
most voice-activated devices have to
be plugged in.
The microphone uses a variety of
unique materials that convert sound
energy directly into electrical energy,
then uses that sound energy to wake
devices from sleep.
It consumes a negligible
amount of power and thats a huge
advantage when it comes to using
devices like hearables, wearables,
smart speakers, and event-detection
systems. Weve been able to extend
battery life by 10 times, Crowley
claimed.
Voice processors
UK based XMOS, which supplies a
number of voice solutions for IoT
products, recently launched the
XVF3000 family of voice processors,
which provide far-field voice capture
using arrays of MEMS microphones.
The company also took the
opportunity to make available the
Above: According VocalFusion Speaker development
to research from kit, which includes an XVF3000
Gartner, there could processor card and a four-mic circular
be as many as 500 microphone array. The kit provides a
smart devices in the quick way to start developing far-field
average home by voice capture applications.
2025. XMOS believes These are the first in a range of
its development kit voice capture products from XMOS,
will help to create said Mark Lippett, CEO. XVF voice
the voice interfaces processors will open up a host of new
needed to control possibilities for designers looking
them to deliver high performance voice
capture.
XVF3000 devices include speech
enhancement algorithms, along with
an adaptive beamformer, which uses
signals from four microphones to
track a talker as they move. There
is also high performance full-duplex,
acoustic echo cancellation.
Crucially, XVF3000 devices can
be integrated with an applications
processor or host PC via either USB,
for data and control, or a combination
of I2S and I2C. That means
developers can quickly add custom
voice and audio processing.
It is becoming increasingly apparent
that the possibilities for voice
Voice will open interfaces are growing dramatically,
up a host of new especially, when according to Gartner,
possibilities for there will be 500 smart devices in
designers looking every home by 2025.
to deliver high Because todays solutions dont
performance scale, there will be a need for a new
voice capture. type of interface and voice looks like
Mark Lippett it could be the new click.
11 July 2017 www.newelectronics.co.uk
www.newelectronics.co.uk 11 July 2017 17
Harnessing the power of
big data
The case for the smart city is well known but the
infrastructure is still evolving. By Neil Tyler.
I
ncreasing urbanisation and growing
populations mean that cities are
growing at rate that many experts
would say is alarming. Within a
matter of just two or three years, 60%
of the worlds population could be
living in an urban environment, with
all the challenges of energy and water
supply, as well as delivering adequate
transport and healthcare.
The way we are shaping and
developing our cities will be critical to
ensuring a sustainable future, says
Patrik Sjstedt, co-head of Hitachis
Social Innovation Business in Europe.
At the same time, we need to make
sure that cities of the future are
about more than finding solutions to
the challenges that exist today. Our
focus should also be on how smart
technology and the power of big data
will help allow us to bring real social
and economic benefits.
Research from Frost & Sullivan
predicts that, by 2025, more than
25 cities around the world could be
described as smart cities, with more
than half of them located in Europe
and North America.
Smart-enhanced surroundings will
become more interactive and capable
of better responding to peoples
needs, suggests Sjstedt. What
needs to be remembered is that
smart cities are not just a concept;
they will make a tangible contribution
to improving society whilst mitigating
the challenges we are currently
facing.
18
The introduction to all elements
of infrastructure of smart, intuitive
technology, capable of autonomously
reporting on and responding to its
surroundings, will be determined by
the availability and ability to process
data in real time. But that data
will not only impact on individuals,
businesses and governments, but
also on how these various entities are
integrated into a bigger framework.
Smart cities will essentially
create an ecosystem which brings
stakeholders together to support
economic development, sustainability,
and environment conservation,
explains Maurizio Colleluori, principal
data engineer, Think Big Analytic.
Big data, mobile technologies and
cloud computing have to be in place
to enable smart cities to function
effectively.
But, says Colleluori, that data systems will generate vast amounts
will need to be cleansed, validated, of it, and, routed through intelligent
profiled, indexed, secured and infrastructure, will help generate
correctly tracked. Crucial to all of this insights.
is data governance, which ensures In the case of the smart city,
that data can be trusted and that those insights will help to address
people can be made accountable for inefficiencies and improve processes.
how data is collected and used. The volume of data being
He also makes the point that, with The way we generated is surging and most of that
a shortage of software engineering are shaping is being sent to the cloud, explains
talent, data management is often left and developing Lee. While it is more cost effective,
Illustration: Adobestock
to inexperienced data engineers. our cities will cloud farms are massive and cant be
Data is at the heart of the smart be critical to housed near a city. They need to be
city concept, says Jackson Lee, VP ensuring a in isolated areas, be cheap to build,
of corporate development in data sustainable have excellent network connectivity
centre services at Colt Technology future and many are looking to run them on
Services. Real-time monitoring Patrik Sjstedt green energy.
11 July 2017 www.newelectronics.co.uk

A good example in Europe of a
data centre that has taken advantage
of lower power costs, abundant green
energy resources, connectivity and
natural cooling is the Lefdal Mine
Datacentre (LMD) in Norway.
Located on the countrys west
coast, the 120,000m2 datacentre
operates exclusively on renewable
energy and is cooled by water from
a nearby fjord. The energy costs
are correspondingly low and the
system achieves a Power Usage
Effectiveness (PUE) of less than 1.15.
Not only can it offer different
technical solutions, but with a
200MW power generation capacity,
it can offer customers a pay as you
grow model.
According to Lee, while data
can be pushed to the cloud and
facilities like LMD address this
growing trend, it is only an effective
solution when data does not need to
be processed in real time.
To process data in real time
and Im talking of mission-critical
or real-time traffic like video, cloud-
based applications or gaming, the
extra latency associated with cloud
computing can degrade performance
significantly and increase security
risks. We are seeing rapid growth in
the development of edge datacentres
www.newelectronics.co.uk 11 July 2017
when users need to manipulate data
in real time without latency.
Typically, edge datacentres
are smaller and, as such, can be
positioned much closer to end users.
They can be located within cities or
the suburbs they are serving, Lee
explains.
Edge data centres provide
speed, he continues. However, the
use a lot of CPUs and, increasingly,
GPUs which means they use a lot of
power. We will also see a growing
trend towards the use of artificial
intelligence (AI) to manage the
sheer amount of data that is being
generated.
The applications that are
developed and deployed will
determine the use of AI, but I can
see AI running a lot of this data, Lee
argues.
The edge is currently dominating
the market but, crucially, the edge
can go even smaller. Edge data
centres could be set up locally,
even in peoples homes. We simply
dont know how much data will be
generated.
Colt will be rolling out its ultra-high
bandwidth Colt IQ network platform
across Europe and the Asia Pacific
this year and this is an important
foundation for its network service
SECTOR FOCUS DATA MANAGEMENT
offerings.
On-going digital transformation
is leading to higher demand for
datacentre capacity as well far more
flexible business and technical
solutions. The edge and cloud offer
different solutions for handling the
data from smarter systems, but I
believe we will see most processing
taking place in the cloud, Lee
suggests.
Smart cities will Smart cities cannot be created
essentially create in isolation and the UK is a hub of
an ecosystem, smart city innovation, with nearly
which brings 100million already spent in
stakeholders preparing cities for a sustainable
together to future. Initiatives like Bristol is
support economic Open and CityVerve in Manchester
development, have sought to encourage greater
sustainability, collaboration to make the concept a
and environment reality.
conservation. The introduction of smart,
Maurizio Colleluori intuitive technology to all elements
of infrastructure, capable of
autonomously reporting on and
responding to its surroundings, will
allow the value of real time data to be
realised, explains Sjstedt. Cities
will only truly become smart when
they can react to all of the data being
generated within them.
Managing digital transformation
has always been a requirement of IT
decision-makers, but the momentum
of technology change is putting a
Left: Located on strain on both data centre capacity
Norways west and the skills and scope of IT teams,
coast the Lefdal as Colleluori noted.
Mine Datacentre has According to Sjstedt, only by
taken advantage combining the efforts of technology,
of low power costs telecoms, governments and
and abundant green businesses will we be able to deliver
energy resources the smart city and data management
is the key to unlocking value from
connected devices across these new
smart city ecosystems.
We have the potential to reduce
emissions, offer better healthcare
and provide more effective transport
services based on the interpretation
and analysis of the relevant data,
he concludes. Real innovation is
required if we are going to make
smart cities a reality.
19

The long haul
Sophisticated truck automation and connectivity solutions exist, but
legislation is needed before they can be used on the road. By Peggy Lee.
W
ith the Google car and
similar initiatives, almost all
the focus in transportation
is on self-driving cars. Trucks,
however, could also benefit from extra
automation and connectivity to enable
preventive maintenance, improve
safety, efficiency and cost. Might we
therefore one day soon see trucks on
our roads with no driver in the front
seat?
Not for at least 10 years; I
dont think the public would accept
it, argued Jrg Rger, president
of commercial vehicle and off-road
at Bosch Mobility. If were talking
autonomous driving for trucks, were
talking about motorways, not inter-
urban traffic. For that, its not realistic
from todays point of view because
when the truck leaves the motorway,
it would need someone to take over.
www.newelectronics.co.uk 11 July 2017
This is because the current level
of technology doesnt provide for this
degree of autonomy yet.
While Europe is in the drivers
seat when it comes to standards for
increasing road safety, legislation
is lagging behind the level of
sophisticated solutions being
developed solutions which could
lead to fully autonomous trucks.
Legislation is currently being
written for stop and go control
systems and turn assist systems,
which Rger believes will appear in
the next couple of years.
WABCO has developed its
own urban turning assist collision
avoidance system, said Dr Thomas
Dieckmann, leader of advanced
development at WABCO. OnCity is a
single-sensor solution and the first to
use LiDAR (light imaging, detection
SECTOR FOCUS SMART
TRANSPORTATION
and ranging) technology. LiDAR
measures distances by illuminating
a target with a pulsed laser light and
measuring the reflected pulses.
The system alerts the driver
visually and acoustically to a potential
collision, both right before and during
a turning manoeuvre, Dr Dieckmann
added. In the future, it will be able
to apply the brake autonomously to
prevent collisions, should the driver
fail to take corrective action.
The next level up in terms of
automation though only currently
envisioned on motorways is
platoons: convoys of trucks linked
and controlled electronically through
short range vehicle-to-vehicle
communications.
Led by the front vehicle, this
communication, coupled with
technologies such as forward
collision avoidance systems, enables
the trucks to accelerate or brake
simultaneously and to follow each
other more closely.
Potential benefits include: better
fuel economy due to reduced air
resistance; reduced congestion;
fewer collisions/accidents; and less
pressure for drivers, who could,
instead of driving, plan routes,
In Volvo Trucks process shipping documents or
Dynamic Steering simply take a break.
system, an electric Several trial platoon runs have
motor is added to been conducted in Europe including
the conventional the European Truck Platooning
hydraulic power Challenge but legislation is still
steering unit wanting.
Platooning needs legislation to
allow trucks to drive at a distance of
eight to 10m, much closer together
than it is currently legal, Rger
explained, and to allow drivers
to be on standby.
There will eventually be
mixed platoons and that needs a
standardised protocol. The expectation
is to see these platoons on the road
between 2023 and 2025.
Not everyone is keen on the
idea however. The Road Haulage
Association believes platooning is not
feasible.
21
 SECTOR FOCUS
Causing queues for vehicles trying
to join and leave the motorway will
simply create even more congestion,
cautioned Rod McKenzie. Of course,
the auto-pilot facility has the ability to
remove human error and mistake
but what happens if the engine goes
wrong?
As drivers wont be concentrating
on the road, they might not be able
to react as quickly when theres a
system failure, for example. Another
downside is that systems could
potentially be hacked.
To resolve these issues, Volvo
Trucks is concentrating on improving
human machine interaction to
ensure good usability. We provide
training and methods for continuous
improvements along with supporting
tools to assist planning and driving,
said Michael Gudmunds, product
manager, soft products.
While the practicality of platoons is
open to debate, commercial vehicles
are already benefitting from increased
connectivity and automation.
Commercial vehicles can
already upload data into the cloud
and download automated software
updates, said Rger.
Infotainment is becoming
more connected with smartphone
integration, voice control and a
Bluetooth hands-free system.
Navigation systems are being used
to optimise engine use and fuel
consumption by providing the driver
with road and environment mapping
information in advance.
Real-time information about road
conditions such as the presence of
black ice can be shared between
trucks via data collected by the
sensors, then processed and stored
in the cloud and retransmitted.
Legislation for advanced
emergency braking systems (AEBS)
and lane departure warning (LDW) has
been in place since 2015.
Commercial vehicle technology
supplier WABCO offers an LDW
system called OnLane, which detects
road markings and vehicle position,
22
SMART TRANSPORTATION
and warn the driver of imminent lane drivers intentions. A principle called
departure via visual, audible and torque overlay corrects unintentional
haptic signals. The camera-based steering movements to keep the truck
solution can distinguish between dead on course.
a deliberate lane change and an Some trucks are already smart,
unintentional drift by identifying the benefitting from advancements in
drivers turn signal usage. automation and connectivity, and
Lane keeping support takes LDW companies are more than eager to
one step further by not only warning take these capabilities to a higher
the driver but also, if no action is A 40tonne level.
taken, by taking steps to ensure the autonomous But there is a long path to full
vehicle stays in its lane. truck frightens autonomy and, as Ruger says: A
Volvo Trucks Dynamic Steering people. 40tonne autonomous truck frightens
which combines conventional Jrg Rger people.
hydraulic power steering with an Autonomous commercial vehicles
electric motor fitted to the steering therefore will not appear tomorrow.
gear also provides lane keeping Development will be steady, Dr
support. Dieckmann concluded. We will see
The electrical control unit an increase in partial automation,
processes data from multiple sensors then further developments in the
and controls the motor at 2000Hz to areas of manoeuvring and highway
work out the trucks direction and the driving.
Remotely-controlled wheel loaders go underground
Smart vehicles are being used for the rst time in mining as part of the Swedish PIMM project Pilot for
Industrial Mobile communication in Mining.
Wheel loaders are used to load the blasted rocks onto trucks. There is always a safety risk for drivers
that unsecured rocks will fall, said senior project manager Eilert Johansson. So better safety is one
benet to controlling heavy machinery remotely from above ground.
According to Johansson, another benet is improved efciency of mining operations as time is no
longer lost bringing people up and down into the mines to change shifts.
To be able to control the wheel loaders remotely as well as for other uses a 4G wireless cellular
communication network has been installed.
We spent a lot of time designing radio wave propagation models for underground, Johansson said.
When youre remote-controlling heavy machinery, you have high bit rate requirements, so need high
bandwidth for the video streams, together with low latency.
As a prototype, a regular wheel loader was used on which the engineers built a remote-control
connectivity solution equipped with six cameras, LIDARs and vibration and audio sensors. Several
technology concepts had to be implemented in the software, such as a safety concept to automatically
stop the machine when communications links are lost.
According to Peter De Bruin, master researcher at Ericsson, 5G components have been included in
the 4G network, with the idea of gradually
updating to 5G.
With 5G, latency will be improved,
capacity to run more vehicles and more
sensors simultaneously will go up, De Bruin
explained. We are planning to add network
slicing and 5G enabled core networks.
Network slicing is a way to realise
different services for different applications
all running on the same network by
creating virtual networks on top of the
physical network.
11 July 2017 www.newelectronics.co.uk
Microbumps: the
probing challenge
Testing 3D-stacked IC technology is challenging as
manufacturers look to retain high compound yields
but solutions are at hand. By Ferenc Fodor.
R
esearch on 3D stacked IC
(3D-SIC) technology has
advanced to the point where
most semiconductor companies
have released or announced 3D-
SIC products. These packages
require multiple chip dies to be
stacked vertically, which results in
dense integration in an ultra-small
footprint, but with benefits in terms of
performance, power and cost.
One challenge in stacking ICs is to
retain a high compound yield and to
not include faulty dice. This requires
testing the dice before stacking them;
for example, through the interconnect
microbumps.
However, many engineers have
considered it impossible to probe
these microbumps because the arrays
are too large and the microbump pitch
of less than 40m is too small.
In response, nanoelectronics
research centre imec has used
National Instruments Semiconductor
Test System (STS) to develop a fully
24
automated system to characterise
prototype probe cards for large-
array, fine-pitch microbumps on
advanced test wafers.
Probing microbumps
Due to its high-precision, 3D-
SIC manufacturing is prone to
defects and every IC needs to undergo
electrical tests to weed out defective
parts and guarantee product quality.
3D-SICs, which typically contain
complex die designs in advanced
technology nodes, need to be tested
using advanced test and design-for-
test approaches.
A number of test challenges are
unique to the stacking process itself.
One is testing dies prior to stacking;
essential to obtain acceptable
compound stack yields and to not lose
good dice in a stack with a faulty die.
The non-bottom dies of the
stacks have their functional access
exclusively through large arrays of
fine-pitch microbumps, which are
Above: The Pyramid too dense for conventional probe
RBI probe card technology.
A common approach is to equip
these dice with dedicated pre-bond
probe pads. However, this comes with
drawbacks, such as increased silicon
area and test application time and a
reduced interconnect performance.
To avoid these drawbacks, imec and
its partners set out to enable probing
directly on the microbumps, a task
previously thought impossible.
The probe challenges for this type
Below: Fully of microbump array include:
automatic test Providing a test system with a
system to evaluate sufcient number of channels for large
microbump probing arrays of microbumps. Test systems
with 1200 channels are usually
expensive and a soft-docking cable
interface with that many channels
would be nearly impossible to connect
and maintain.
Landing with all probe tips on the
microbumps. Advanced probe cards
are non-see-through, they require
an advanced probe station that
determines probe-tip landing locations
using software overlay of microbump
images from a downward-looking
camera and images of the probe
11 July 2017 www.newelectronics.co.uk

tips from an upward-looking camera.
Sufficient positioning accuracy is
required from both the probe tips with
regard to each other, as well as from
the probe station while stepping over
the various dice on the wafer.
Establishing a proper electrical
contact between probe tip and
microbump. The quality of the probe
contact is expressed by the contact
resistance (Rc). This not also depends
on the metallurgy of the microbumps
and the shape and material of the
probe tips, but also on the force with
which the probe station presses the
wafer against the probe tips and the
cleanliness of the probe tips.
Probe marks should not impact the
stacks interconnect yield negatively.
The prime function of microbumps is
to form a reliable electrical connection
between two stacked dice. Marks
should not deteriorate the interconnect
stacking yield and probe marks should
be limited in topology, or be erasable
through a reflow process.
The approach should be economically
feasible. Pre-bond test is constrained
by cost: it should not cost more to
perform the test than to skip it.
To address these challenges, imec
worked with probe card supplier
Cascade Microtech, which provided
prototypes of its advanced Pyramid
Rocking Beam Interposer (RBI) probe
cards. These probe cards contain an
IC design specific probe core, which
includes a thin film with MEMS-type
probe tips. Cascades probe cores
www.newelectronics.co.uk 11 July 2017
SYSTEM DESIGN SYSTEM TEST
Left: The heel of the parametric and functional tests,
tip makes physical while rack 2 is dedicated to
contact with the microbump probing and contains
wafer a PXI-4072 digital multimeter
connected to an ultra-wide switch
matrix with nine concatenated
PXIe-2535 modules, each with
136 output columns. This made
it possible to connect each of the
four DMM channels under software
control to any of the 1224 SMX
output columns.
The system supports two-
and four-point Kelvin resistance
measurements between any pair of
support more than 1200 core I/Os, probe tips for daisy chains as
sufficient for WIO1 (JEDECs JESD-229 well as between a single probe tip
standard for Wide I/O single data rate and all other probe tips ganged, for
DRAM). The RBI probe tips require characterisation of that single probe
less than 1gf/tip to make proper tip when all probe tips are shorted
electrical contact and the heel of the through the probed wafer.
tip makes physical contact with the
wafer in such a way that the probe Conclusion
mark is typically only 6 1m. Tests in 300mm wafers designed and
To prove the feasibility of manufactured in-house by imec and
microbump probing with these probe featuring microbumps with various
cards, imec built a fully automatic test metallurgies, pitches, diameters
system consisting of a dual CM300 and sizes), demonstrated that all
probe station from Cascade Microtech, WIO1 probe tips landed on the
a hard-docking STS test head with PXI corresponding microbumps; the actual
test instruments from NI, a test head contact resistance between probe tip
manipulator from Reid-Ashman and and microbump was approximately
test programme and data analysis 0.1. However, the resistance of the
software based on LabVIEW and trace through the thin film membrane
developed at imec. Author prole: on the probe core is often included
The NI STS test head contains Ferenc Fodor is a in the measurement and this is
two PXI racks with test instruments. test development approximately 5.
Rack 1 holds instruments for engineer with imec The probe marks on the Cu were
small and, while the probe marks
on Sn were larger, they could be
Benets working with NI products removed through reflow. Imec also
demonstrated there were no stacking
NI STS hard-docking test head enabled imec to make
interconnect yield differences
more than 1200 instrument-to-probe connections,
between all four cases of bottom/top
something not possible with a manual cable interface.
microbumps probed/not-probed.
The PXIe-2535 switch matrix modules allowed
Through cost modelling, it was
concatenation of an ultra-wide switch matrix with an
possible to demonstrate that, for
unlimited lifetime of FET switches. Its unlimited simultaneous
single-site testing, the Pyramid
cross-point connection provided the ability to perform probe
Probe cards, although expensive,
checks.
outperformed pre-bond testing through
The test times were short; about 100 two-point resistor
dedicated probe pads financially.
measurements per second.
Crucially, the work showed that
LabVIEW provided an easy-to-program interface to the
microbump probing was technically
system.
and economically feasible.
25
 EMBEDDED DESIGN POWER

L
ithium-ion batteries require
considerable care if they are to
operate reliably over a long period.
Because they cannot be operated
to the extreme end of their state
of charge (SOC) and because their
capacity diminishes and diverges
over time and with use, every cell in a
system must be managed to keep it
within a constrained SOC.
Providing sufficient power for a
vehicle requires tens or hundreds
of battery cells to be configured in
a long series, generating 1000V or
more. The battery electronics must
not only operate at this high voltage
and reject common mode voltage
effects, but also measure and control
each cell in these strings differentially.
The electronics must also be able to
communicate information from each
cell in a battery stack to a central
point for processing.
Operating a battery stack in
a vehicle or in other high-power
applications imposes tough
conditions, including working in

The power to manage

an environment with significant
electrical noise and wide operating
temperatures.
While the battery management
electronics are expected to maximise
operating range, lifetime, safety and
reliability, car makers will also expect
them to minimise cost, size and weight.
Battery monitoring
Linear Technology announced its first
high performance multicell battery
stack monitor the LTC6802 in
2008. The device could monitor up to
12 Li-ion cells within 13ms and with
0.25% maximum total measurement.
Many LTC6802 ICs could also be
connected in series to enable the
simultaneous monitoring of every cell
in high voltage battery strings.
The latest addition to this part of
Linears portfolio is the LTC6811,
a multicell battery stack monitor
incorporating an ultrastable voltage
reference, high voltage multiplexers
and dual 16bit delta-sigma A/D
converters. An LTC6811 can monitor up
Battery management systems continue the drive towards higher
reliability in EVs and HVs. By Greg Zimmer.
to 12 battery cells connected in series
with a measurement accuracy better
than 0.04%. Using the devices fastest
A/D converter mode, all cells can be
measured within 290s. Meanwhile,
with eight programmable third order low
pass filter settings, the LTC6811 can
also undertake noise reduction.
Each LTC6811 includes two 1MHz
serial interfaces: an SPI interface for
connecting to a local microprocessor;
and the proprietary two wire isoSPI
interface. The isoSPI interface
provides two communication options:
multiple devices can be connected in
a daisy chain to the BMS master (host
processor); or multiple devices can be
connected and addressed in parallel
to the BMS master.
Modular battery packs
To accommodate the large number
of cells required for high powered
automotive systems, batteries are
often divided into packs, then dist
ributed throughout available spaces in
the vehicle. Typically comprising up to
24 cells, modules can be assembled
Author prole: in different configurations to suit
Greg Zimmer is multiple vehicle platforms. A modular
senior product design not only simplifies maintenance
marketing and warranty issues, but can also be
engineer, signal used as the basis for very large battery
conditioning stacks. This allows battery packs to be
products, distributed over larger areas, for more
with Linear effective use of space.
Technology, In order to support a distributed,
part of Analog modular topology within the high
Devices. electromagnetic interference

www.newelectronics.co.uk 11 July 2017 27

 EMBEDDED DESIGN POWER

environment of an electric or mechanical failure of connectors,

hybrid vehicle (EV/HEV), a robust cables and wiring harnesses in high-
Cell 15
communication system is required. + C2 vibration automotive environments.
Cell 14
The isolated CAN bus and Linears + C1 V Cable To date, the metal and high-EMI
Cell 13
isoSPI both offer road-proven solutions (twisted surroundings in vehicles were thought
+ pair)
for interconnecting modules in this Cell 12 to be too harsh for a wireless system to
+ C11 C12 isoSPI
Cell 11
environment. + C10 operate reliably. However, SmartMesh
Cell 10
While the CAN bus provides a well- + C9 networking offers a redundant
Cell 9
+ C8
established network for interconnecting Cell 8 interconnect system through its use of

battery modules in automotive
applications, it requires a number of
additional components. For example,
implementing an isolated CAN bus via
the LTC6811s SPI interface requires
the addition of a CAN transceiver, a
microprocessor and an isolator. The
primary downside of a CAN bus is the
added cost and board space required
for these additional elements.
An alternative to a CAN bus
interface is Linears two wire isoSPI
interface, which is integrated into
every LTC6811 (see fig 1). This
interface needs only a simple
transformer and a single twisted pair,
rather than the four wires required
by the CAN bus. The isoSPI interface
also provides a high RF noise immune
interface, allowing modules to be
daisychained over long cable lengths
and to support data rates of up to
1Mbit/s.
Wireless BMS
In a wireless battery management
system (BMS), each module is
interconnected via a wireless
connection, instead of a CAN bus
LTC6811-1
+ C7
Cell 7
+ C6
Cell 6
+ C5 Cable
Cell 5
+ C4 (twisted
Cell 4 pair)
+ C3
Cell 3 isoSPI
+ C2
Cell 2
+ C1 V
Cell 1
cable or an isoSPI twisted pair.
Linear is currently demonstrating a
wireless automotive BMS in a BMW
i3 concept car. Within this concept
car, an LTC6811 battery stack
monitor and SmartMesh wireless
mesh networking products replace the
traditional wired connections between
the battery packs and the BMS. The
demonstration of a fully wireless
BMS car represents a significant
breakthrough that offers the potential
for improved reliability, lower cost and
reduced wiring complexity for large
multicell battery stacks for EVs/HVs.
Automakers are challenged to
ensure the driving public that electric
and hybrid/electric vehicles are not
only safe, but also reliable. Linear
Technology is looking beyond the
safety and reliability of the battery
monitoring IC to address the potential
path and frequency diversity, allowing
BMS master
wireless messages to be routed around
isoSPI
obstacles and to mitigate interference.
(LTC6820)
MCU
Field-proven in industrial Internet
of Things applications, SmartMesh
embedded wireless networks support
highly reliable data transmission in
Figure 1: Linears harsh environments. By delivering the
isoSPI interface reliability of wires, while eliminating
can be used to mechanical connector failures, the
daisychain battery wireless BMS concept car shows that
stack monitors wireless technology has the potential
to improve overall system reliability
significantly and to simplify the design
of automotive battery management
systems.
Additional benets
A BMS with a SmartMesh network has
the potential to provide functionality
currently unavailable in a wired
system. The wireless mesh network
enables the flexible placement of
Battery battery modules and makes possible
monitoring and the installation of sensors in locations
additional sensor previously unsuitable for a wiring
interconnections harness. Additional data informing the
using SmartMesh accuracy of battery SOC calculations
wireless networking such as current and temperature
may be collected by the BMS
Master by simply adding SmartMesh
enabled sensors. SmartMesh time
synchronises each node automatically
to within a few microseconds, and
timestamps measurements at
each node accurately, allowing the
battery SOC and state of health to be
determined more precisely.
A SmartMesh node, with local
processing at each module, improves
normal BMS operation and offers the
potential for smart battery modules,
where module diagnostics and
communication may be available to
enhance assembly and service.

28 11 July 2017 www.newelectronics.co.uk

T
wo years ago, ARM decided to
investigate the way in which
its teams undertook hardware
design. Using tools developed for
cloud computing, the company now
sifts through every result from the
thousands of tests and simulations it
runs on a daily basis in order to look
for patterns.
RTL verification generates a lot
of data, Bryan Dickman, director
of engineering analytics in ARMs
technology services group, told the
TV&S Verification Futures seminar
earlier in 2017. Its becoming a big-
data problem. One of the questions
which ARM wants to answer is are
we running dumb cycles?.
Dickman noted: We do a lot of
analytics around bugs. We drill into
Learning by
example
bugs by methodology and areas of
the design: what methodologies
find bugs and when. It is interesting
to see how you are finding bugs
against the amount of verification
you are running. You can start
asking questions like I have been
running lots of cycles, but no bugs
were discovered: were they dumb Machine learning techniques could play a leading role in getting bug
cycles?.
So far, the emphasis has been free chips to market more quickly. By Chris Edwards.
on building visualisation tools to
help engineers analyse the patterns. ML, most think its neural network- to a collection of data points, to
The next step is to build tools that related. But the data doesnt have to more advanced techniques, such as
can learn from the data. We are be modelled by a neural network. support vector machines that, before
looking at how we can take machine- Singapore-based Plunify combines the advent of deep learning, proved
learning algorithms and apply them domain knowledge with some forms successful in image recognition.
to the data, then design predictive of ML in its InTime tool, which tunes Earlier this year, tool supplier
workflows that improve productivity, the parameters used to compile FPGA Solido Design Automation launched
Dickman said. designs for performance and area. its ML Labs initiative. The company
Machine learning (ML) has started Plunify co-founder Kirvy Teo noted: built a toolkit of techniques for use
to take hold across the EDA world in The nature of ML in this area is in its own tools, such as one that
general and more than 20% of papers different from the things we see at We are looking analyses models of the low-level
at the 2017 Design Automation Google. Neural networks work very at how we can behaviour of the cell libraries used
Conference (DAC) dealt with ML well for lots of data points, but we take machine in IC design on advanced processes.
techniques. But the applications are are dealing with thousands of data learning In these processes, statistical
not going to be distributed evenly. points, rather than billions. algorithms and variation at the nanometre scale
Access to data is one of the biggest Fellow co-founder and CEO apply them to strongly affects performance and the
issues. HarnHua Ng added: We cant tell the data, then effects of temperature and voltage
Image:Adobestock

Speaking at a panel session
during DAC, Ting Ku, senior director
of engineering at nVidia, said ML
techniques need to be tuned to the
problem. When people talk about
customers First youve got to run ten
thousand compilations.
EDA tool developers can apply
techniques ranging from fairly simple
regression algorithms, that fit a curve
design predictive are not always linear. Traditionally,
workows this demands many simulation
that improve runs to characterise the cells at all
productivity. possible operating points. The tools
Bryan Dickman analyse how the variables in existing

30 11 July 2017 www.newelectronics.co.uk


libraries interact and build models to
predict the parameters for a subset
of simulations that will create an
effective library.
Amit Gupta, Solido CEO, explained:
We have been focused on ML for
variation-aware design for the past
seven years and are expanding that
into other areas. The first area is
characterisation, but we are getting
demand from customers to expand
into other areas within EDA.
Sorin Dobre, senior director of
technology at Qualcomm, points to an
expansion of ML across more of the
tasks needed at the physical-design
level. At 10nm and 7nm, we see
a lot of process variation. You have
to validate across multiple process
corners.
The problem that faces physical-
verification teams is the number
of combinations of temperature,
transistor speeds and voltage that
require checking under brute-force
methods. The question is to how
to get good quality results without
an explosion in process corners.
ML can bring a 10x improvement in
productivity.
One area where RTL verification
teams have seen success with ML
methods is in eliminating some of the
dumb verification cycles that concern
ARMs designers. Harry Foster, chief
scientist at Mentor, now a subsidiary
of Siemens, says. Can I minimise
the number of tests and still achieve
With novelty detection
Wit
Test coverage (%)
www.newelectronics.co.uk
030_NELE_JUL11.indd 31
RESEARCH & DEVELOPMENT
a reasonable level of coverage?
There are papers on that and it is a
technique now being applied in the
industry.
Yield analysis is another area
where ML is having an impact. It can
predict the types of geometry in a
dense design that are likely to lead to
lithographic or other process failures
and thus dud chips. Improving
yield by two or three points has a
significant economic benefit, Dobre
notes.
Machine learning should make it
easier to build simplified models of
complex interactions that defy rule-
based design methods. For example,
researchers at Georgia Institute of
Technology proposed using models
generated automatically from
electromagnetic measurements and
simulations to help organise devices
and I/O pins on multi chip modules
so that signals do not suffer from
crosstalk or IR drop.
Where ML doesnt work well is
where you have to guarantee finding
something, like manufacturing test.
With verification, I can get away with
less than 100% coverage, because
my intent is mitigating risk. I cant
with manufacturing test, because it
leads to returns, Foster said.
The next step is to try to reuse
the learning across projects. This is
one of the obstacles that slows down
adoption of data mining and ML in
RTL verification.
Without novelty detection
>19hrs of server farm time
3000 6000
Number of applied tests
11 July 2017
EDA
Courtesy Cadence
Researchers at Foster continued: The issue with
Georgia Tech, [RTL] verification is that although
amongst others, are people say we have lots of data,
looking at whether the reality is we dont. If Im starting
machine learning a project, I have no data. How can
can be used to we be successful in applying ML
build better heat- techniques if our data is limited?
aware models that It cant be based just on pure data
will improve chip- mining; it has to incorporate some
package codesign domain knowledge.
Dobre says the issue of lack
of data is reflected in physical
implementation. The question is
how can you maintain cycle times
when you go to the next process
technology? We have highly
automated flows from RTL to gates,
but they are not repetitive they are
based on experience. If you bring in
a new design team, you start from a
pretty low level of experience.
One possibility is to use data
mining to look at repetitive tasks,
such as timing closure, and to use
ML to automate some of them, while
learning about the new processes as
each project continues. We need to
capture designer knowledge during
A machine-learning implementation and reflect it back
algorithm developed to the design community and the
by the University foundry side to fix issues, Dobre
of California at argued, adding this would need data
Santa Barbara and from many sources.
Freescale lters There is a need for flexibility from
achieved the same the major EDA companies to work
coverage in 310 with these solutions, he concluded.
tests as a normal If you have a closed system and
testbench did in dont allow anyone to interact with
6010. your tools, we cant do anything.
31
03/07/2017 14:26
032_NELE_JUL11 03/07/2017 10:46 Page 1