Ede 1112

01 november 2012
PLC-Compatible PSU
near-Field Communications
Hot Topic p. Hot T

Survival
Guide
TheelecTronica
Make the
Most of
Munichs
Mega-
techfest.
TABLEOFCONTENTS
38. Hot Topic
NFC: Dont Leave Home Without It
52. Design Ideas
Automotive Power-Conditioning Circuit Eliminates Power-Hold
Relay; more
56. Applications
Design Your Own PLC-Compatible PSU
60. Advertiser Index
01.11.12 ELECTRONIC DESIGN EUROPE
P. 4
P. 8
4. Editorial
Can The Chip Makers Beat The Eurozone Blues?
8. Product News
Intermediate Bus Converters Incorporate Digital Control; more
18. Power Design
Tighten The Gap Between ADC Datasheet Specs And The Real World
24. Technology
The Electronica 2012 Survival Guide
3 ELECTRONIC DESIGN EUROPE
CAN THE CHIP
MAKERS BEAT
THE EUROZONE
BLUES?
A
big question on peoples minds at
this years Electronica show in Mu-
nich will be how the international
electronics industry is riding the
nancial storm created by the euro-
zone crisis.
Europes semiconductor manufacturers
have all experienced a drop in revenues
this year. As an industry group, the decline
is around 8% with some companies experi-
EDITORIAL
By Paul Whytock Editor-in-Chief
4 01.11.12 ELECTRONIC DESIGN EUROPE
ELECTRONIC DESIGN EUROPE
encing a 15% reduction. In
contrast to those depress-
ing gures, industry analysts
and pundits are saying 2013
and 2014 will be positive
with growth rates estimated
between 6% and 9%.
So given those encour-
aging predictions, answer-
ing how the semiconductor
industry will fare relative to
the eurozone cash dilem-
ma is easy. Or is it?
Hardening Attitudes
Europes politicians still
havent hammered out a
strategy that resolves the
problems created by a
debt-laden boom in Eu-
rope that led to the bust.
Many senior government
ministers claim it wasnt
foreseen, which is hard to
believe given past industry
examples like the boom
and bust that deeply affect-
ed the electronics industry
back in 2000. As it stands,
huge monetary bailouts to
countries that have created
overwhelming debt arent
providing an instant nan-
cial solution.
The crisis is aggravated
by the hardening attitudes
of the voting population
of countries that have to
shoulder the major bur-
den of providing bailout
cash to save the Euro. A
recent opinion poll in Ger-
many found that only 25%
of Germans think Greece
should stay in the eurozone
or get more nancial help.
So against this backdrop,
what are the prospects for
international semiconduc-
tor companies?
Amid rising economic
concernsincluding the
Euro zone crisis, slowing
manufacturing growth in
China, and stubbornly high
unemployment in the Unit-
ed Statessecond-quarter
growth for the global semi-
conductor industry was
highly disappointing, says
Dale Ford, senior director
of electronics and semi-
conductor research at in-
dustry analysts IHS.
Approximately two-thirds
of the worlds semicon-
ductor suppliers saw their
revenues decline in the
second quarter of 2012
6
compared to the same period in 2011.
This weak performance bodes ill for the
semiconductor industrys growth pros-
pects for 2012, Ford continues.
Brighter Prospect
But what about next year? World Semi-
conductor Trade Statistics maintains the
worldwide semiconductor market will
grow by 7.2% in 2013 and reach a total
value of 255 billion, followed by 4.4%
growth to reach 266 billion in 2014.
And there are some signs that this opti-
mism could be well placed.
Among the recent positive news about
Europes semiconductor business, the
Taiwan Semiconductor Manufacturing
Company (TSMC) has decided to invest
1.11 billion in Netherlands-based lithog-
raphy specialist ASML Holding NV. By do-
ing this, TSMC has joined Intel in taking a
stake in Europes largest chip-equipment
maker to secure future technology.
TSMC gets a 5% equity holding in
ASML for 834 million of the total in-
vestment, and 276 million will go on
research and development of next-gen-
eration lithography technologies. TSMC
also will gain access to machines under
development that will reduce manu-
facturing costs. Intel agreed to invest
as much as $4.1 billion last month in
ASMLs industry investment programme
to fund innovation.
So how are some of Europes major
chip companies doing in terms of equity
value? ARM has stormed into the lead
when it comes to market capitalisation
and is way ahead of Inneon, STMicro-
electronics, and NXP, making it Europes
most valuable semiconductor company.
In capitalisation terms, the gures are
Arm at $12 billion, NXP at $5.8 billion,
and STMicroelectronics at $4.8 billion.
Looking Further Aeld
International electronics, IT, and com-
munications companies are recognising
the need to fully develop and nancially
exploit emerging markets. Many CEOs
will admit they have implemented this
business strategy but do concede that
a lot of work remains if the full nancial
trading potential of such territories is to
be engaged.
A recent report from Global Intelligence
Alliance (GIA) concluded that technology
and telecom companies are expecting
50% of their global revenues will come
from emerging markets by 2017. This is
an enormous gure and one that I treat
with caution given the uneasy political
and nancial characteristics of some
territories. However, GIA makes clear in
its report that Brazil, Russia, India, and
China are still the top four most important
emerging markets for 2012-2017.
Of these, GIA maintains that Russia
is least favoured, with most companies
focusing more on India, Brazil, and Chi-
na. Some of the larger technology com-
panies are targeting Brazil as their top
emerging market and surprisingly favour
Vietnam over Russia as their fourth most
preferred market.
Also seen as worthwhile are Indonesia,
Mexico, South Africa, the Philippines,
Argentina, Thailand, Turkey, Malaysia,
Singapore, and the United Arab Emir-
ates. Thats quite a mix of countries, and
it will be interesting to see future trading
trends for electronics companies that do
push into those emerging markets.
So it would appear that despite the eu-
rozone crisis, there remains ample op-
portunity for the semiconductor industry
to move forward in 2013. Sectors such
as automotive, medical, and industrial
electronics continue to provide a stable
source of revenue for the chip makers.
But its the erce market competition be-
tween giant adversaries like Apple and
Samsung in the smart-phone and tablet
markets that provide the semiconductor
industry with very high-volume orders.
Analysts IHS recently reported on how
Apple is the worlds leading OEM pur-
chaser of semiconductors. This year it is
expected to buy nearly $28 billion worth
of semiconductors, up 15% from $24
billion in 2011. Apple is also expected
to achieve the strongest growth in chip
spending among the worlds top 10 OEM
semiconductor buyers and is set to ex-
pand its lead in global chip purchasing
in 2013 with growth of over 12%.
Having big spenders like that sounds
like a great prospect for the semicon-
ductor companies and their share-
holders, but with one very important
proviso: beware an overly dominant
purchasing position held by one global
OEM customer. Chip suppliers could
see immense pressure to cut pricing.
Its far better to have a broad selection
of protable OEMs needing semicon-
ductor products that are prepared to
pay a fair price. Q
ELECTRONIC DESIGN EUROPE 7
Tualatin, Ore., U.S.: CUI has released two
series of fully regulated intermediate bus
dc-dc converters following its licensing
of Ericssons FRIDA II digitally controlled
advanced bus converter portfolio. The
NQB and NEB series, which are part of
CUIs Novum Advanced Power range,
are pin-compatible with Ericssons
BMR456 and BMR457 series, respec-
tively (see the gure).
The quarter-brick NQB and eighth-
brick NEB series are based on a 32-
bit ARM MCU with power-optimising
rmware designed to maximise and
atten the efciency curve, providing
peak performance
across a much
wider range of
loading conditions
than other
intermediate bus
converters.
With typical
efciencies as
high as 96.4%
and optimised
efciency from
20% to 100% load,
the modules suit
applications where
loading conditions
can vary greatly.
The rmware is
also designed to handle input voltage
transients with slew rates of up to 0.5
V/s, while keeping the output voltage
within 10% and ensuring the output
voltage doesnt trigger over-voltage
protection.
The NEB and NQB series both provide
a digital interface that complies with PM-
Bus 1.2. A range of parameters can be
set via a USER_STORE memory block.
A number of protection features can be
set, including over-temperature, over-
current, and over-voltage.
The converters are available in two
input range congurations: 40- to 60-V
PRODUCT NEWS
By Staff
INTERMEDIATE
BUS CONVERTERS
INCORPORATE
DIGITAL CONTROL
CUIs NQB and NEB series of fully regulated intermediate bus dc-dc converters are pin-
compatible with Ericssons BMR456 and BMR457 series, respectively.
dc input for datacom applications and
36 to 75 V dc for telecom applications.
The NEB eighth-brick series is offered
with power ratings ranging from 264
to 300 W, while the NQB quarter-brick
series is available in 420- to 468-W
congurations.
Output voltage for all models is
preset at 12 V dc. For dynamic bus
applications, users can dynamically
adjust the NEBs output voltage via
PMBus from 6.9 to 13.2 V dc and the
NQBs output voltage from 4.0 to 13.2
V dc. For higher-power requirements,
optional Droop Load Sharing is offered
in both series.
CUI
www.cui.com
CSLOC Ecosystem For Video
Surveillance Over Coax Grows
Milpitas, Calif., U.S.: Intersil has expanded
its Techwell SLOC (Security Link Over
Coax) standardisation ecosystem, de-
signed to accelerate the global adoption
and implementation of hybrid IP-based
security systems for the video surveil-
lance market (see the gure). The com-
pany has joined forces with nearly 100
others, including Sony, Altronix, and
Hikvision, to address the need for in-
creased security in next-generation vid-
eo surveillance systems.
SLOC is a modem physical-layer
technology that enables high-resolution
megapixel and HD IP-based cameras
to operate within the existing analogue
CCTV coaxial cable infrastructure with-
out the need for new cabling. It consists
of a camera transmitter and a DVR/NVR
receiver, a combination that en-
ables high-performance and ro-
bust bidirectional communication
between the camera and the video
recorder over long cable runs.
SLOC supports up to 500 m of
coaxial cabling, eliminating repeat-
ers. Embedded analogue CVBS
(color, video, blanking, sync) video
features enable latency-free live
viewing and control of digital IP
cameras.
Intersil
www.intersil.com
330-W Open-Frame Supply
Achieves 15-W/in.
3
Density
Simi Valley, Calif., U.S.: N2Powers
330-W XL330-54 ac-dc power
supply offers an impressive com-
bination of power and density. The
open-frame power supply delivers
a power density of 15 W/in.
3
with
only 13CFM cooling. Measuring 34 mm
(1.35 in.) high, its 76- by 135-mm (3 by
5.3 in.) footprint makes it 32% smaller
than similarly rated 1U power supplies,
which tend to come in 102- by 152-mm
(6 by 4 in.) form factors (see the gure).
Achieved through synchronous recti-
cation, Or-ing MOSFETs, and a unique
active power factor correction (PFC)
design, the XL330-54s 90% maximum
full-load efciency yields 23% less heat
than most comparable power supplies,
Intersil has expanded its Techwell SLOC
standardisation ecosystem, which is designed to
accelerate the global adoption and implementation
of hybrid IP-based security systems for the video
surveillance market.
www.micreI.com
2011 MicreI, Inc. AII rights reserved.
MicreI, Innovation Through TechnoIogy and HyperLight Load are registered trademarks of MicreI, Inc.
MLF and 0LFURLeadFrame are registered trademarks of Amkor TechnoIogy, Inc.
Extend Your Battery Lifel
with Micrel's "Green Power" Regulators
HyperLight Load
8ynchronous Buck
IdeaI for use in:
Mobile handsets
Portable media/MP3 players
Portable navigation devices (GPS)
WiFi/WiMax/WiBro modules
Solid State Drives/Memory
Wireless LAN cards
Portable applications
MI C2315 5 d e l i v e r s e x t r e me l y h i g h e f f i c i e n c y f o r
por t abl e pr oduct s and gr een home/ of f i ce appl i ances.
Mi c r el ' s Hy per Li ght Load
Regul at or s s ave pr ec i ous

power over t he ent i r e l oad r ange f or ' Gr een' Devi ces.
The MC23155 features internal MOSFETs able to deliver up
to 2A output current while consuming just 23A of quiescent
current inside an ultra small 2.5mm x 2.5mm Thin MLF
package.
For more information, contact your local Micrel sales representative
or visit Micrel at: www.micrel.com/index.php/en/products/power-
management-ics/dc-dc-switching-regulators-/step-down-internal-
switches/article/7-mic23155.html
Ultra High Efhciency - > 93%
Ultra 8mall 8ize - 2.5 x 2.5mm
Ultra Powerful - 2A
Ultra Fast - 4MHz
MIC23155
0.47H
2.2F
SS
PGOOD
EN
4.7F
VOUT
Down to 0.7V
VIN
2.7V to 5.5V
2.5mm x 2.5mm
Thin MLF
12 01.11.12 ElEctronic DEsign EuropE

minimising cooling requirements and
increasing system reliability.
The XL330-54 has two primary out-
puts, 54 V/6.1 A and 12 V/9 A, in addi-
tion to a 12-V/1-A auxiliary output. The
54-V output is 2250 V dc isolated from
the other outputs, which makes it com-
patible with Power-over-Ethernet (PoE).
Furthermore, the design allows active
current sharing on the 54-V output for up
to four power supplies to be operated in
parallel. N+1 operation can be confg-
ured as well.
Other features include universal ac
input (90 264 V ac) and active inrush
current protection of 10.5 A
at 240 V ac, which does not
change during operation.
N2Power
www.n2power.com
MEMS-Based Sensor
Enables Smart Gas
Meters
Kyoto, Japan, and Geneva, Swit-
zerland: Omron and STMicro-
electronics have teamed up
to develop a MEMS-based
(microelectromechanical sys-
tems) gas-fow sensor with
unusual built-in correction for
differences in gas composi-
tion (see the fgure). The sen-
sor is intended primarily for
gas meters, which are mov-
ing toward smart metering solutions like
electricity meters. The sensor will enable
higher precision and reliability, accord-
ing to the companies.
The sensor combines Omrons
MEMS thermal fow transducer with
STs high-performance analogue front-
end IC, delivering high-precision gas-
fow-rate measurement with excellent
reproducibility. Gas meters built
around the Omron/ST solution wont
need to be confgured for a certain
type of gas at the time of shipment
or installation, as they are intrinsically
compensated for both temperature and
N2Powers 330-W XL330-54 ac-dc power supply delivers
a power density of 15 W/in.
3
with only 13CFM cooling.
Its 76- by 135-mm footprint makes it 32% smaller than
similarly rated 1U power supplies.
13 ElEctronic DEsign EuropE
pressure variations and a built-in circuit
compensates for the variation of multiple
gas composition. The sensor is dust-
resistant to comply with international gas
meter standards.
STMicroelectronics
www.st.com
Omron
http://components.omron.eu
Power-Aware Tools Complement
Low-Power 32-Bit MCU
Austin, Texas, U.S.: Silicon Labs
has reduced the power con-
sumption of its latest 32-bit
MCU family to 175 A/MHz in
active mode and less than 250
nA in sleep mode with the real-
time clock (RTC) enabled at 3.6
V (see the fgure).
Part of the SiM3L1xx family,
which is based on an ARM Cor-
tex-M3 core operating at up to
50 MHz, the Precision32 mixed-
signal MCUs include various
power-saving peripheral and
architectural innovations that
can reduce current consump-
tion below that of many 8-bit
MCUs. They also come with the
companys new power-aware
development tools.
Active-mode power reduction
is achieved through a number
of innovations. For example,
dynamic voltage scaling adjusts the
internal device voltage in response to
changing conditions. An integrated,
high-effciency dc-dc buck converter
reduces active-mode power. Dedicated
peripherals such as a data transfer man-
ager, Advanced Encryption Standard
(AES) encryption block, and run-time
encoder accelerate the processing of
RF protocol for wireless applications
without CPU intervention, reducing sys-
Omron and STMicroelectronics
have teamed up to develop a
MEMS-based gas-flow sensor with unusual built-
in correction for differences in gas composition.
Designed for smart metering solutions, the sensor
will enable higher precision and reliability.
tem power. Enhanced direct memory ac-
cess (DMA) can reduce protocol-related
power by 90%.
On-chip peripherals have also been
optimised to reduce sleep-mode power.
The charge pump generates a power-
effcient input voltage for the device
circuits in sleep mode, which reduces
analogue sleep currents by 35% and
digital sleep currents by 50%. Also, the
SiM3L1xx MCUs support a multi-alarm
RTC for clocking and interrupts and a
sleep-mode UART for low-power device
communication. An integrated sensor
interface provides sensor stimulus and
measurement while the MCU is in sleep
Silicon Labs has reduced the power consumption of its SiM3L1xx 32-bit MCU family to 175
A/MHz in active mode and less than 250 nA in sleep mode with the RTC enabled at 3.6 V. The
MCUs also come with the companys new power-aware development tools.
15 ElEctronic DEsign EuropE
mode. The autonomous sensor interface
continues to count in sleep mode and
can wake the MCU after a count over-
fow or when the count reaches a pro-
grammable threshold.
Silicon Labs has also introduced two
capabilities for lowering power con-
sumption to its complimentary Eclipse-
based integrated development environ-
ment (IDE) and AppBuilder software for
Precision32 MCUs. Power Estimator pro-
vides a pie chart of total supply current
and additive currents for enabled periph-
erals to show where power is being con-
sumed, while Power Tips provides soft-
ware confguration guidance that helps
minimise current consumption.
Silicon Labs
www.silabs.com
Baseplate-Cooled Power
Supplies Boast 90% Efficiency
In Harsh Environments
Fyfeld, U.K.: XP Powers CCH series of
compact single-output baseplate-cooled
400-W and 600-W ac-dc power sup-
plies is intended for applications in
harsh environments (see the fgure). Up
to 90% effcient, the CCH supplies par-
ticularly suit sealed-box applications,
where all heat-generating components
are attached to the baseplate to allow
heat dissipation through the sealed-box
chassis or heatsink. No forced-air cool-
ing is required.
Measuring just 214 by 102 by 43 mm
(8.43 by 4.02 by 1.69 in.), the CCH
supplies are up to 50% smaller than
other baseplate-cooled products of a
similar power level. Their high effciency
may also mean a smaller heatsink
can be specifed, reducing the overall
footprint required.
Both the 400-W CCH400 and the
600-W CCH600 series accommo-
date the full universal input range of
90 to 264 V ac without derating, suit-
ing them for designs used worldwide.
These single-output units are available
with the popular nominal output volt-
ages of +12, +24, +28, or +48 V dc.
Compliance to MIL-STD-461 for
emissions and MIL-STD-810 for shock
and vibration ensure the CCH series
fts use in commercial off-the-shelf
(COTS) military applications in addi-
tion to a broad range of industrial and
commercial equipment used in out-
door and harsh environments.
The units also comply with the inter-
nationally recognised safety standard
EN/UL 60950-1 for IT equipment
and meet EN55022 level B conducted
and level A radiated emissions.
A fully featured signal set includes
remote on/off, remote sense, ac
OK, and overtemperature warning
and shutdown. A +5-V dc standby
output is also provided. The supplies
current-sharing capability allows for
parallel sharing of the load across up
to three CCH units. The CCH series
has a wide operating temperature
range of 40C to 70C, with a maxi-
mum baseplate temperature of 85C.
n
XP Power
www.xppower.com
Offering up to 90% efficiency, XP Powers CCH single-output baseplate-cooled 400-W and
600-W ac-dc power supplies are intended for harsh environment applications.
W
A
N
T

T
O

B
E

P
A
R
T

O
F

T
H
E

C
O
M
M
U
N
IT
Y
?
W
ith the m
ost content and reach of any design
focused publication in the electronics m
arket,
Electronic Design brings you the latest professional
inform
ation through cutting-edge m
ultim
edia
publishing.
Check out what our m
em
bers are saying:
I like to read Electronic Design to stay aware
of technology advances in com
ponents and
equipm
ent, says Hugh Howes.
Very, very useful. I feel confident of being able
to stay up-to-date with the latest technology and
associated devices through reading Electronic
Design, says W
illiam
Brooks.
Join today and see w
hat everyone is
raving about!
To subscribe, visit:
http://electronicdesign.com/subscribe
or call 1-866-505-7173.
C
onnect w
ith us:
W H E R E T E C H N O L O G Y C O M E S F I R S T
TIGHTEN THE GAP
BETWEEN ADC
DATASHEET SPECS
AND THE REAL WORLD
POWER DESIGN By Michael Steffes Intersil Corp.
Its doubtful that a design will fully realize an ADCs datasheet
numbers, but applying SNR and SFDR optimization techniques
help it inch closer to the ideal.
A
ll high-speed analog-to-digital
converter (ADC) datasheets show
specied performance that really
should be considered an ideal
response under the best of all
possible test conditions. Most systems
cant use coherent sampling, super-narrow
bandpass ltering, or very high insertion-loss
lters just before the ADC.
Upon realizing that those ADC datasheet
numbers cant be achieved in real systems,
designers then must take an asymptotic
18
approach. This can involve analytic
tools, but will still probably require some
iteration on the nal interstage lter
design to trade off the conuence of
numerous issues.
High-speed ADC suppliers move
mountains to extract the best signal-
to-noise ratio (SNR) and spurious-free
dynamic range (SFDR) performance in
lab and automatic test equipment (ATE)
tests for their converters. A very real
arms race exists among chip suppliers
to achieve the best numbers for their
respective datasheets.
Nevertheless, its highly unlikely
that any real system will reach those
numbers. Therefore, the best designers
can hope for is to not throw away
any more dynamic range than is
absolutely necessary. To do that, one
must have intimate knowledge of the
unusually constrained
ADC characterization
environment, along
with the techniques for
holding onto as much of
that intrinsic capability
as possible in an actual
product.
In The Lab
High-speed ADC manu-
facturers have all evolved
similar methodologies for
device characterization. Most of these
methodologies are in the form of single-
tone or multi-tone fast Fourier transform
(FFT). The key results collapse down to
a few numbers: SNR, SFDR, harmonic
distortion (HD), and equivalent number
of bits (ENOB). At the most basic level,
pulling the best SNR out of an ADC typi-
cally requires:
Extremely low phase-noise source and
clock
High clock-signal amplitude to get
through the sampling threshold as fast
as possible
Coherent sampling, which involves
phase-locking the clock to the source
and digitizing an integer number of
input cycles; this removes the need
for windowing in the FFT, eliminating
spectral leakage
Its doubtful that a design will fully realize an ADCs datasheet
numbers, but applying SNR and SFDR optimization techniques
help it inch closer to the ideal.
Unbalanced
input
Differential
output
50-, 3-dB pad
T1 T2
1.41:1
13.7
13.7
20
Driving the ADC from very low source
impedance, which helps minimize
Johnson noise from the source and
often improves SNR for unbuffered
ADC inputs
In actuality, there are more factors
at play here. Looking at a typical ADC
characterization board, its clear that
most suppliers use permutations on a
two-input transformer structure (Fig. 1).
With the two 13.7- resistors and the
1.41:1 turns ratio of T2, the circuits ter-
mination impedance referred to the input
is very close to the desired 50 when
the ADC input resistance is placed in
parallel with the 27.4 of shunt imped-
ance.
This structure is very useful in
attaining a low driving impedance (12.5
, in parallel with 13.7 , equals 6.5
on each side). However, it needs a
higher input power level to reach full
scale (FS). As such, combining the
3-dB pad, the transformer insertion
losses, and the step-down for the
2-VPP-FS ADC requires 17 dBm (or 4.5
VPP) at the input.
The Real Secret
Although all of this effort achieves
the desired 105-MHz input SNR for the
ISLA216P25 ADC (up to 74.5 dBFS),
it still requires one more super-critical
bit of test hardware thats seldom
shown in evaluation board schematics.
Figure 1s passive interface is very
broadband. Looking at the transformer
specications, it offers a 3-dB span of
10 MHz to more than 1 GHz. With such
a wide analog bandwidth, the test circuit
would integrate lots of noise at the input
pins, which degrades SNR.
To avoid the noise problem, ADC
manufacturers employ narrow band-
pass lters ahead of the input (Fig. 2).
This (expensive) tunable bandpass l-
ter is extremely effective at constraining
the spectrum at the input to the circuit
in Figure 1. Thus, it delivers very low
integrated noise and vanishingly small
harmonic distortion terms for single-
tone tests.
Tunable lters become vital
components when trying to obtain the
highest SNR datasheet numbers for
high-performance ADCs. However,
its unrealistic to expect designers to
include their equivalent in any real
system. Also, these 5%-of-center-
frequency narrow-width lters pass a
narrow spectrum to these very high-
clock-rate ADCs.
Its acceptable to characterize the IC
with a less than 5% bandpass-ltered
test signal. Still, most real systems
need to use a much larger portion of
the available Nyquist band (FS/2). By
using real lters with wider passbands,
the integrated spot noise at the ADC
input will be greater than that delivered
during ADC characterization. The SNR
achieved in production systems must be
less than the ADC datasheet numbers.
Approaching ADC-Only Numbers
Another specication integral in op-
timizing ADC performance is SFDR in
the FFT that processes the ADCs raw
output. The harmonic-distortion terms
at the inputs of the ADC combine in-
phase with the ADC-generated terms.
For instance, if the lab-tested ADC gen-
erates a harmonic with
amplitude of 85 dBc,
and the input signal
level also is 85 dBc,
the FFT is expected to
show 79 dBc.
To minimize SFDR
degradation from the
datasheet ADC-only
tests (using the ex-
tremely clean signal
coming out of the lters
in Figure 2), the input distortion levels
should probably be greater than 20
dBc below the rated ADC performance.
Achieving super high linearity (i.e., low
distortion) implies exceptionally high
quiescent power in the nal-stage ampli-
er and/or a very high loop gain in that
stage for the intended frequency band
of interest.
Such low harmonic distortion signals
are only possible if the last stage ampli-
er has a bandwidth far wider than the
desired signal band. As a result, a noise-
power, bandwidth-limiting, inter-stage lter
is required to reduce the SNR degradation
2. Before the test signal
reaches the circuit in Figure
1, it goes through a tunable
bandpass filter. The lower
filter is used for the signal
path, and the upper handles
the clock signal.
introduced by that ampliers broadband
noise.
Final system performance mostly
comes down to the passive lter
between the last amplier and the ADC.
Some design considerations must be
applied:
The test circuit in Figure 1 added
7-dB insertion loss from the input to
the ADC. But its usually preferable
to target 1.5- to 2.5-dB insertion-loss
lters to keep the required voltage
swing out of the last stage amplier to
a range only slightly higher than the
intended ADC input range.
Conicts arise when keeping the lter
input impedance R to the amplier
outputs relatively high (to avoid
unnecessary distortion degradation
due to heavy loading of the last
amplier). Designers also must include
the ADC input impedance as part of
the lter-element design. Because
these specs often have no specied
tolerance, its best to include external
elements that combine with the ADC
terms so they dominate the value for
the lter design.
Relatively low impedance values can
limit the noise added by the resistors
and present a low source impedance
to the ADC.
The passive lter elements should
display reasonable tolerance.
Mismatches in the series element
values for differential lters will
lead to converting some part of the
output differential signal to common
mode. (Also, if theres a common-
mode component to the output, it will
get converted to differential.) This
matching is intended for the series
elements. Therefore, the differential
capacitor elements often can exhibit
5% tolerance, while good results have
been achieved using 1% resistor and
2% inductor tolerances.
There are some tradeoffs obviously,
and the nal design often becomes a
bit empirical. For further insight, the
ISLA112P50/55210EV1Z evaluation
daughterboard offers a detailed ex-
ample design (www.intersil.com/en/
products/data-converters/high-speed-a-
d-converters/ISLA112P50-55210EV1Z.
html). A 12-bit, 500-Msample/s ADC
experienced only a 0.6-dB SNR degra-
dation, while SFDR from 1 to 6 dBc
(depending on the frequencies). Q
MICHAEL STEFFES is the senior applica-
tions manager, High Speed Signal Path, at
Intersil. Previously, he was the market de-
velopment manager/high-speed products
at Texas Instruments and had worked at
Burr-Brown and ComLinear. He has a BSEE
from the University of Kansas and an MBA
from Colorado State University.
THIS MONTH IN ELECTRONIC DESIGN
Dont miss out on the best of the rest that Electronic Design has to offer. The November 6, 2012 issue of the North
American edition of Electronic Design offers additional articles that will benet you and your design work. And theyre all
online now!

ENGINEERING FEATURE
TABLETS ECLIPSE THE PC MARKET
BILL WONG EMBEDDED/SYSTEMS/SOFTWARE EDITOR
TECHNOLOGY REPORT
FREE DOWNLOADABLE SPICE TOOLS CAPTURE AND SIMULATE ANALOG CIRCUITS
DON TUITE ANALOG/POWER EDITOR
ENGINEERING ESSENTIALS
THE FUNDAMENTALS OF SHORT-RANGE WIRELESS TECHNOLOGIES
LOUIS E. FRENZEL COMMUNICATIONS EDITOR
DESIGN SOLUTION
REDUCE LOSSES IN RF SCHOTTKY-PIN LIMITER CIRCUITS
CHIN-LEON KIM AVAGO TECHNOLOGIES
IDEAS FOR DESIGN
INEXPENSIVE SOLUTION PROTECTS SENSITIVE DEVICES
FROM SURGES
GIRISH CHOUDANKAR EMPHATEC INC.
EDITORIAL
WHAT OS IS IN YOUR SMART PHONE OR TABLET?
JOE DESPOSITO EDITOR IN CHIEF
LAB BENCH
IS KICKSTARTER THE NEW WAY TO GET CAPITAL?
BILL WONG EMBEDDED/SYSTEMS/EDITOR
1 2
3
R
cs
R
sl
PDLY
ADLY
SBUS
V
ref
SPRG
DPRG
RLEB
Ct
SS
UVLO
U1
LTC3722-1
OutA
OutB
CS
OutD
OutC
OutE
OutF
Comp
FB
V
CC
V
sec
V
lter
V
load
R
load
C
lter
Gnd PGnd
D1
D2
TR
TC
M1
Bridge
1 turn
TECHNOLOGY

2012
SURVIVAL
GUIDE
Make the most of
Munichs mega-techfest.
THE
ELECTRONICA
E
very two
years, the in-
ternational
electronics
community
converges on
the Bavarian city of Munich to ex-
hibit at or visit the worlds largest
electronics show. This year is no
exception with an estimated 2600
exhibitors and 75,000 attendees
expected at Electronica 2012 at
the New Munich Trade Fair Centre.
Visitors will see some of the lat-
est innovations in semiconductor
technology, displays, nano-sys-
tems, sensors, test and measure-
ment, EDA, passive components,
electromechanics, power elec-
tronics, printed-circuit boards
(PCBs), and automotive and wire-
less technologies.
But this event is about more than
the electronics on show. Elec-
tronica also is host to forums and
conferences covering a multitude
of technologies.
For example, automotive broad-
band data connectivity started
with 3G. Now, 4G (LTE/LTE Ad-
25
By Paul Whytock Editor-in-Chief
Be prepared. November weather in
Munich can range from bright and sunny
to cold and snowy.
vanced) is available in the United
States and partly in Europe. One
presentation during Electronica
2012 will examine 4G technology
for automotive applications, look-
ing at deployment worldwide with
an outlook on the importance
and application of broadband
data connectivity for tomorrows
vehicles.
Energy Issues
Smart energy technology, one
of the major themes of the ex-
hibition, will be put under the
microscope. Manufacturers will
present solutions and products
for energy efciency, energy stor-
age, LEDs, and smart grids. In
fact, the future of power grids will
be a key subject.
Against the backdrop of the
global transition to alternative en-
ergy sources and the expansion
of power grids, Electronica 2012
will present technical solutions
for the intelligent supply of power
in the future. Leading internation-
The New Munich Trade Fair Centre is
huge, so plan your trip to Electronica
2012 to make the most efficient use of
your time.
al executives will also discuss this topic,
Semiconductor Solutions for the Smart
Grid Challenge, at the CEO Roundta-
ble, moderated by Killian Reichert.
Renewable energies
are playing a growing role
in the supply of energy.
Given the increased inte-
gration of solar and wind
energy, energy produc-
tion is becoming increas-
ingly decentralised. Grids
will have to be expanded
and controlled and moni-
tored more intelligently.
Electronica 2012 will
present technical solu-
tions that will secure the
supply of energy in the
future.
Embedding The Right
Technology
Whether its smart
phones or industrial con-
trol systems, selecting the
appropriate processors,
operating system, drivers, and network-
ing technology is imperative. But unless
all the elements in an embedded sys-
tem interact correctly, the whole design
will have a limited function within a nar-
row range of applications.
The conference is our way of ad-
dressing this challenge for embedded
system engineers who can come here
to gather specic information about se-
lecting a platform that is right for them,
TECHNOLOGIES AT
ELECTRONICA 2012
There will be more than 130
forum and conference events at
Electronica 2012, covering:
Power electronics
LEDs and innovative light
concepts
Microsystems technology
Micros & DSPs
Smart metering
ARM-based computers &
processors
Industrial control &
communication
Embedded computing
Energy management & efficiency
Small form-factor boards: tools &
software
Electronic lighting software
development
Medical electronics
Organic electronics
Learn more about Electronica 2012
at www.electronica.de.
said Nicole Schmitt, Electronica exhibi-
tion director, commenting on the shows
Embedded Platforms Conference.
Getting Onboard
All too often, PCB technology can be
overshadowed by the more glamorous
technical developments in the IC sector.
In close cooperation with the Electrical
and Electronics Manufacturers Associa-
tion (ZEVI), the PCB Marketplace com-
munications platform will make its debut
at Electronica 2012. Its lectures will be
in Hall C1, while its exhibits will be in
Halls B1 and C1.
Getting To The Show
To see and learn about all this tech-
nology, youve got to get to the show.
With nearly 75,000 people in atten-
31
dance, getting there can be a chal-
lenge, especially since nearly half of
the people visiting Electronica 2012
will come from abroad.
If youre ying into Munich, you have
a few choices for getting to the exhibi-
tion grounds. Unless youre in a group,
taxis are very expensive. The buses are
the best and most economical trans-
port, running from just outside the air-
port terminal.
If your rst port of call is central Mu-
nich, use the S Bahn trains or the Luf-
thansa City Bus Service, which runs into
central Munich and terminates at the
Hauptbahnhof (Munichs main railway
station). Most people get to the Messes-
tadt (show grounds) by U Bahn.
Terror Of The Tickets
Munich has a fantastic train system
thats a joy to use. Its clean, reliable,
and efcient. The S Bahn lines are the
suburban trains that run within the city
and to outlying areas. The U Bahn con-
stitutes the central city lines. You need
to get the U2 line to the show. Theres
just one major problemthe mind-
numbingly complex ticket tariffs that
bafe overseas visors and Munich citi-
zens alike.
And here we based the software on David
Beckhams left foot.
The Munich rail system works on an
honesty policy. There are no entrance
or exit barriers, so nothing is stopping
you from travelling without a ticket. Dont
attempt to do so, though, no matter how
tempting and easy it looks. Decipher
the complex instructions on the ticket is-
suing machines and get your ticket.
Before entering the platforms, make
certain you use one of the small ma-
chines nearby to date and time-stamp
your ticket before boarding the train.
Forget to do this at your own peril. If
you do, youre considered to be riding
without a ticket. Munichs trains have
plain-clothed ofcials riding on them,
and they spot-check passengers to see
if they have a ticket.
And dont think that by acting like a
dumb non-German-speaking tourist
Electronic Designs Electronica 2012 Event Site
http://electronicdesign.com/electronica2012
youll get away without a ticket when
the ticket inspectors who regularly pa-
trol the trains question you. You wont.
Theyve heard it all before, and the nes
are steep. You can nd transportation
help at:
http://europeforvisitors.com/munich/
articles/u-bahn-s-bahn.htm
www.muenchen.de/verticals/Trafc_
Transport/227515/index.html
You also can nd a printable map of
the Munich rail system at www.mvv-
muenchen.de/leadmin/media/Dateien/
plaene_en/pdf/netz_2012_en.pdf.
Let The Evening Fun Begin
Most of Electronicas international visi-
tors stay in Munich. But dont be tempt-
ed to stay in a hotel near the show. Its a
very boring thing to do and youll miss
out on a fantastic city. Need convinc-
ing? Take a look at www.panorama-cit-
ies.net/munich/images_munich.html.
Instead, stay as near to Marienplatz
the central square of Munichas you
can. From here, its an easy train ride
into the show in the morning. But forget
that for the moment. In the evenings,
its time for beer and food. Munich is
world famous for its variety and quality
of beer, and a great pub guide can be
found at www.europeanbeerguide.net/
munipubs.htm.
I prefer wine to lager beer, but when
Im in Munich, I drink both! The beer is
great, and so are some of the beer halls
PRIMARY INDUSTRY SECTORS
ON SHOW
Industrial electronics 41%
Automotive 22%
Medical electronics 16%
Electronics for other
consumer goods
16%
Power engineering 13%
Wireless 13%
Military electronics 12%
Wired telecommunications 12%
Railway technology 9%
Aviation/aerospace
technology
8%
Consumer electronics 7%
Office equipment, IT (OEM) 7%
Specialty retail, retail,
distribution
4%
Services 4%
Banking/insurance <1%
where you drink it. Theyre invariably
linked to the breweries that produce the
stuff, so theyre a must visit. To get a taste
of which breweries are the major players,
go to www.beerdrinkersguide.com/BDG-
Website/MunichBeer/BigSix.htm.
The Hoff Brau House is the best
known. Consequently, its too much of
a tourist trap for my taste. But its where
Adolf Hitler made some of his speeches
prior to World War II, so it does hold
Marienplatz, which is the centre of Munich, is a
good place to start your evening exploration of
the city.
a certain fascination. Or if you like the
idea of dining in Hitlers favourite res-
taurant, go to the Osteria Italiana at
Schellingstrasse 62.
But back to the beer halls. One of
my favourites is the Augustiner Gro-
gaststtte. Its huge, historic, authentic,
brash, and very Bavarian, and both
the beer and food are good. Dont ex-
pect microscopic portions of designer
food. The beer comes in bucket-sized
glasses, and the food is served in Fred
Flintstone portions. The suckling pig
is particularly good. The Augustiner
is centrally located and only a short
stumble from Marienplatz. Take a look at
www.augustiner-restaurant.com.
Yes, your superglue
does set quickly.
TECHNOLOGY AREAS VISITED
Semiconductors 50%
Passive components 40%
Test and measurement 40%
Displays 34%
Sensors 34%
Embedded systems 31%
Power suppplies 31%
PCBs, other ciruit carriers, EMS 29%
Electronica automotive 28%
Electromechanics/system
peripherals
28%
Electronic wireless 27%
Electronics design (ED/EDA) 24%
System components 19%
Electronica micro/nano-systems 10%
Information and services 7%
You may prefer a restaurant that doesnt want to
serve you mountainous joints of pig accompanied
by bucketloads of sauerkraut and dumplings you
could play soccer with. Fortunately, Munich has
plenty of ne dining experiences like the restau-
rant at the Knigshof Hotel (www.koenigshof-hotel.
de/en/restaurants). It isnt cheap, but it boasts one
of the best wine cellars in the city.
Alternatively, theres the well-rated Spaten-
haus with its splendid views of the Munich Opera
House. There are plenty of excellent restaurants
in this city, and a lot of them can be found at
www.munich-info.de/restaurants/lists/restaurants_
en.html.
Finally, rst-time visitors to Munich should try the
weisswurst. These white sausages are a tradition-
al Bavarian specialty rst created around 1850.
Theyre made from veal and bacon, avoured with
parsley onions, ginger, and cardamom. They look
anemic and maybe a little unattractive. I like them,
but you should only eat them between breakfast
and noon.
The sausages are made daily, and the people
of Munich eat them in the morning because the
meat is unsmoked. Without modern refrigeration,
they would perish the same day. Munichs citi-
zens still eat them for breakfast with a substantial
dollop of sweet mustard and very often a cold
glass of Pils. And, dont eat the skin. Only eat the
meat inside.
Guten appetit, and have a great show. Q
The Mnchner weisswurst traditionally is eaten before
noon. Its an acquired taste. Try it Bavarian style with a
glass of cold beer.
HOT TOPICS
S
oon, smart phones
will include near-eld
communications (NFC)
wireless technology so
they can be used like
keys or credit cards.
Users simply wave their phone
near an NFC reader, or tap
it, and the devices exchange
data to make a transaction.
Automatic paring is another
emerging application.
The Radio Technology
NFCs maximum range is
about 20 cm with a typical use-
ful range of 4 to 5 cm, which
benets security. The technol-
ogy utilizes the near eld rather
than the more familiar far eld
(see Whats The Difference Be-
tween The EM Near Field And
The Far Field? at electronicde-
sign.com).
The far eld comprises the
orthogonal electric and mag-
netic elds that extend out from
the antenna beyond several
39
By Louis E. Frenzel Communications Editor
Near-eld communications, an extremely
short-range wireless technology, will
soon revolutionize payment and access
systems.
NFC
DONT LEAVE HOME
WITHOUT IT
NFC
DONT LEAVE HOME
WITHOUT IT
A tap on the reader with an NFC-
enabled smart phone automatically
pays for the fare on this train. (courtesy
of NXP Semiconductors)
wavelengths. The waves behave as
Maxwells equations predict where the
electric and magnetic elds exchange
energy and rejuvenate one another
along the signal path. The eld strength
decreases with distance (d) by a factor
of 1/d
2
.
The near eld is within one wavelength
or less of the antenna. It also consists of
an electric and magnetic eld, although
the magnetic eld is more dominant.
The signal strength drops off by a factor
of 1/d
6
, making it far less useful.
Essentially, the near eld is the mag-
netic eld produced by the transmit an-
tenna. It can be considered the primary
of an air-core transformer, and the re-
ceive antenna can be seen as the sec-
ondary winding of that transformer. The
attenuation makes the overall effective
wireless range very short.
NFC operates on the unlicensed
13.56-MHz frequency. It is authorized
in Parts 15 and 18 of CFR 47 from the
Federal Communications Commission
(FCC). Most other countries have autho-
rized it as well. The modulation is ampli-
tude shift keying (ASK) with either 10%
or 100% modulation. The transmitted
binary data uses either the Manchester
code or a modied Miller code to ensure
transmission reliability.
a.
b.
Smart phone (target)
Cellular NFC Tx/Rx
Wi-Fi Bluetooth GPS
Cellular NFC Tx/Rx
Wi-Fi Bluetooth GPS
Active
Active
Active
Near eld
Near eld
Reader (initiator)
Smart phone (initiator)
NFC Tx/Rx
Network
Passive
Tag/sticker (target)
Flash
memory
Self-power
NFC Tx/Rx
MCU
1. NFC uses both active and passive devices. For example, a reader/initiator would poll a smart
phone/target, both in active mode (a). Or, a smart phone would be the initiator polling a tag/sticker/
target (b).
a.
Bit time
Binary 1
+V
+V
1 1 1 1 0
0
0
0 0 0
Binary 0
Binary 1
Binary 0
0 V
NRZ-L
b.
Binary 1: +V
Binary 0: V
c.
Manchester (ac)
Manchester (dc)
Modilied Miller d.
2. Binary waveforms used in NFC include the standard
NRZ-L format (a), an ac Manchester coded wave (b), a
dc Manchester code (c), and a modified Miller code (d).
The Manchester and Miller codes provide more reliable
reading and easy clock recovery.
Data rates are 106, 212, or 424
kbits/s depending on the cod-
ing and modulation percentage.
Some NFC devices use standard
NRZ-L encoding. Binary phase
shift keying (BPSK) modulation
is an alternative at the 106-kbit/s
data rate. Typical signal band-
width is 7 kHz or up to 1.8 MHz
depending on the data encoding
and speed.
A data rate of 848 kbits/s is also
available in some devices, though
it is not part of the approved
standards. A faster very high bit
rate (VHBR) modication to the
standard is under consideration
and will boost the rate to 6.8
Mbits/s for some applications.
NFC Modes, Devices, And Protocols
There are two basic operating
modes for NFC: active and pas-
sive (Fig. 1). In the active mode,
a battery or power supply fully
powers both communicating de-
vices. In the passive mode, one
of the communicating devices is
fully powered while the other is
fully passive. The passive device,
called a tag, derives its dc oper-
ating power from the receive RF
signal transmitted by the active
device.
Radio frequency identication (RFID)
tags work the same way. The passive
device powers up and transmits data
back to the active device using load
modulation, which is a form of ASK with
a low modulation percentage. Load
modulation has the data modulate an
848-kHz subcarrier that in turn modu-
lates the main 13.56-MHz carrier. As a
result, the signal varies the impedance
of the listening device, which translates
to a form of ASK.
Both the active and passive devices
use Manchester coded 10% modulated
ASK for both 212- and 424-kbit/s
operation. Active devices utilize the
Miller code and 100% modulated ASK
for a 106-kbit/s rate to ensure an initial
connection. Figure 2 illustrates the
standard NRZ-L code and the NFC
coding options.
The basic communications mode is
half duplex, where one device transmits
at a time while the other receives. The
operation is listen before talk. One
of the devices is an initiator that must
listen on the channel and transmits only
if no other signal is present. The initiator
polls the other devices that may come
near it. The other device, the target,
listens and then responds to the initiator
according to a formal protocol.
Other modes of operation include
read/write, peer to peer, and card emu-
lation. Read/write operations are em-
ployed to transmit data from one device
to another with both active and passive
devices. The initiator either reads or
writes to a passive device.
In peer-to-peer mode, two ac-
tive devices exchange data to
establish a link for other trans-
missions. The card emulation
mode works like an active de-
vice reading a passive device
like a smart credit card or tag.
The multiple NFC standards,
designated NFC-A, NFC-B, and
NFC-F, dene several slightly
different transmission technolo-
gies. Each species a different
data rate, modulation, coding,
or operational mode (see the ta-
ble). The initiator polling device
Record A Record B Record C
Header Payload
Identier Type Length
3. Also called a message, the NFC Data Exchange Format
(NDEF) comprises one or more records. A record includes
the desired payload plus a header comprising identifier,
length, and payload fields.
attempts to detect the specic mode of
the responding device and then cong-
ures itself to the appropriate technology
to complete the transaction.
Also, the NFC standards dene four
basic types of passive tags from Type
1 to Type 4. Each type has a different
memory capacity and matches one of
the popular standards. Types 1 and
2 have 96- and 48-byte to 2-kbyte
maximum storage and operate at 106
kbits/s. Types 3 and 4 run at 212 or 424
kbits/s and have either 1 Mbyte or 32
kbytes maximum, respectively.
The NFC standards further specify a
message encapsulation format called
the NFC Data Exchange Format (NDEF)
to use in the normal course of operation.
Each transmission is called a message,
and each message comprises one or
more records (Fig. 3). A record includes
the desired payload plus a dening
header, which has identier, length, and
payload type elds. The payload is typi-
cally a URL or a data type dened by
the standard NFC Record Type Deni-
tion (RTD) le.
NFC Standards
Most basic NFC standards were de-
rived from RFID and smart card stan-
dards. They have become formal Inter-
national Organization for Standardization/
International Electrotechnical Commis-
sion (ISO/IEC) standards, including those
standards originally developed by par-
ticipating companies:
ISO/IEC 14443A (NXP, formerly Philips
MIFARE)
ISO/IEC 14443B (Inneon)
JIS X6319-4 (Sony FeliCa)
The RF NFC standard is ECMA 340
(European Association for Standardizing
Information and Communications
Systems). It is designated NFCIP-1 or
Near Field Communication Interface and
Protocol. ISO/IEC adopted it as 18092.
There is also NFCIP-2, called ECMA
352, and ISO/IEC 23917.
The NFC Forum (www.nfc-forum.
com), a non-prot promotional group
of companies, establishes and main-
tains a wide range of related specica-
tions and standards related to NFC. It
also provides testing and certication
programs to promote interoperability of
NFC devices. EMVCo, a joint business
venture of Europay, MasterCard, Ameri-
can Express, and Visa, manages and
maintains specications for smart cards,
point of sale (POS) terminals, ATMs, and
related devices.
NFC Security
If NFC is to be used in lieu of credit
card payments or for access to criti-
cal facilities, the
transmitted data
must be secure.
NFC has inher-
ent security sim-
ply because its
very short range
prevents signals
from traveling
too far. But that
doesnt mean it
cant be hacked.
A high-gain di-
rectional antenna
and sensitive
receiver could
eavesdrop on
NFC signals,
although the
hacking receiver
setup may not
be that incon-
spicuous.
Security risks also come from other
forms of hacking. For example, data
corruption could occur when false
data is transmitted to an NFC reader or
other enabled device. Data also can be
modied during transmission.
During man in the middle attacks,
hackers access the transmitted data
and change it before retransmitting it.
These attacks arent likely, but they are
possible. The best way to protect the
data from these forms of corruption is
to encrypt or otherwise use techniques
to secure the radio channel. Virtually all
NFC radios are encrypted.
Although NFC standards are already
well dened and developed, additional
progress is needed in point-of-sale in-
frastructure integration to fully realize
the technologys potential, explains
Ron Vetter, IEEE Computer Society
member and founder of Mobile Educa-
tion LLC. Because mobile payments
will likely be the big driver for NFC, ad-
AS3911
SPI
XTO XTI
Oscillator Bias
CSI
CSO
RFO1
RFO2
RFI1
RFI2
Regulators
Automatic
antenna tuner
External
eld detector
Phase and
amplitude
detector
VHBR receiver
Capacitive
sensor
1-W VHBR
transmitter
Analog-to-digital
converter
Logic
Wakeup timer
Level shifters
RC oscillator
4. The austriamicrosystems AS3911 features dedicated control logic
implementing all the basic NFC standards for a reader. It uses an external
27.12-MHz crystal, has an SPI external interface, and complies with the
faster 6.8-Mbit/s VHBR standard. Best of all, its 1-W transmitter eliminates
the need for an external power amplifier.
45
dressing customer concerns with secu-
rity and privacy will also play an impor-
tant role in how rapidly the technology
is adopted.
NFC Applications
NFC has many potential uses. How-
ever, the primary target application is
mobile payments. Instead of using a
credit card, customers at stores and
other venues will use their NFC-em-
bedded smart phone as the payment
device. Users will be able to pay for
goods and services at restaurants, re-
tail stores, parking lots, theaters, sports
stadiums, and specialty shops as well
as on buses, trains, taxis, and perhaps
even airlines worldwide with just a
swipe of their phone.
Access is another potential use. Peo-
ple who are authorized to enter secure
buildings, facilities, and other areas will
use their smart phone as an electronic
key. NFC also could be used to access
homes, car doors, and computers.
Simple data exchange is another pos-
sibile NFC application. The peer-to-peer
mode would allow phones or other de-
vices to exchange data. Business card
information could be transferred be-
tween phones. Additionally, data could
be transmitted between laptops and
printers. Data rates are too slow for vid-
eo or digital camera data transfers, but
the potential will be there as higher rates
become available.
Pairing, which is the process of get-
ting two wireless devices talking with
each other, is one of the more promis-
ing uses for NFC. Its usually necessary
in establishing communication in Wi-Fi
and Bluetooth systems. By incorporat-
ing NFC in these devices with pairing
drivers, the connections would take
place automatically without user inter-
actions. Nothing is more aggravating
than having to pair two radios before
use. NFC is a real solution to this mad-
dening problem.
Reading tags or smart stickers is an-
other interesting application. Inexpen-
sive read-only tags can be placed on
almost any item so a smart phone can
read it. The tags may provide a URL
for additional information, or they may
store text and graphics for advertising.
Promotional posters could provide en-
hanced data from a tag. Maps could
be accessed. Tags on products for sale
could provide specications, features,
and price information.
The Mobile Payment Option
The industry appears to be focused
on making NFC the key to cardless
payments, but it will take more than the
seamless wireless link. The credit card
companies, banks, cellular carriers, and
retailers will have to cooperatively as-
semble a massive system that will make
these payments possible. That system is
slowly beginning to emerge.
In fact, there are multiple efforts to
build an e-commerce system that all
can use. A single standard does not
seem possible. The goal is to capture
all of those millions and billions of
transactions and dollars and get a
piece of the action. The current mobile
payment effort is all about making
money and controlling that ow of funds.
Collaboration among players results
in new organizations and systems. It
appears as though several will coexist in
the e-payment realm.
The rst major payment system was
Googles Wallet. This cooperative effort
between Google, MasterCard, and
Citigroup emerged over a year ago, but
it hasnt been widely used. There are
more Google Android phones with NFC
than any other type of phone, so some
activity has occurred.
Another major effort, Isis, is a
collaboration of the cellular carriers
including AT&T, Verizon, and T-Mobile.
Trials are expected to occur in Salt Lake
City, Utah, and Austin, Texas, anytime
now. The Merchant Customer Exchange
amalgamates large retailers like Wal-
Mart, Target, 7-Eleven, Best Buy, CVS,
Lowes, Royal Dutch Shell, Sears, and
Sunoco. Other major vendors like
Apple, Amazon, and Microsoft have not
announced e-payment options.
More than 140 separate e-payment
initiatives are currently in play. Most wont
survive or will operate only in a narrow
sphere of inuence. The larger efforts will
pay off eventually. In addition to garnering
a part of the revenue, most of these
ventures also want to gather customer
data for market research and to enrich
themselves with targeted ads, coupons,
and other sales efforts.
The smart card, an alternative to NFC
payment, has been around for many
years, and its use is continuing to grow.
Smart cards have a built-in chip with
a processor and read/write interface
that interacts with readers at retail
and restaurant sites. Contacts on the
embedded chip let the reader connect
to the card. The smart card is more
secure than traditional magnetic strip
credit cards.
The major standard for such smart
cards is EVM, which is a joint venture
of Europay, MasterCard, and Visa. The
organization managing the EVM effort
is called EVMCo. American Express,
JCB International, MasterCard, and Visa
jointly own EVMCo. Most popular credit
cards use the EVM standard, which is
based on existing ISO/IEC standards
7816 for contact cards and ISO/IEC
14443 for contactless radio-frequency
identication (RFID) chips.
Today there are numerous trials on
NFC for mobile payment and other mo-
bile commerce activities that are being
held by mobile network operators as
well as by players like Google, PayPal,
and retailers like Starbucks. New play-
ers like Target, Wal-Mart, Best Buy, and
CVS are also looking to launch NFC
payment services, says Jagdish Re-
bello, director of consumer and commu-
nications at IHS iSuppli.
These trials are aiming to raise cus-
tomer awareness for the technology, im-
prove the user interface, and sort out the
5. Based on the Microsoft Windows Phone 7
platform, the Nokia Lumina 610 smart phone
uses INSIDE Secures MicroRead v.3.4 NFC
controller and Open NFC protocol stack
software.
business model as various nodes in the
value chain seek to prot from NFC and
develop new offerings that take advan-
tage of the technology. As these issues
work themselves out, NFC is well poised
to be a key enabling technology in the
near future, Rebello says.
Designing With NFC
Adding NFC to a smart phone, de-
signing an NFC reader terminal, or
specifying an NFC tag is a relatively
simple process. The biggest challenge
lies with the phone designs as the NFC
represents one more radio that must
be added to an already impressive col-
lection of wireless devices typically in-
cluding several cellular radios, Wi-Fi,
Bluetooth, GPS, and even FM in some
cases.
Finding the space is the key problem
as the low 13.56-MHz frequency
requires larger components. The chips
are tiny, but the antenna is a printed-
circuit board (PCB) loop or an inductor
on a ferrite core that must be tuned
and matched to the chip. This requires
the most space. The extra power
consumption may also be a problem in
some designs.
The design boils down to selecting a
chip and squeezing it and the antenna
into available space. Large retail reader
designs are easier since they offer the
most space and can use improved
antennas as well as ac supplies.
Multiple vendors offer chips. Most of
the larger semiconductor manufacturers
have some NFC component including
RFID tags. The leader in this space is
NXP (formerly Philips Semiconductor,
one of the founders of NFC), with an
estimated 80% market share.
One of the most widely used NFC
devices is the NXP PN65K. This
two-chip module includes the NFC
transceiver and controller combined
with a Secure Smart Card controller
for security. The transceivers 8051
microcontroller complies with all
NFC active and passive modes and
standards. The two chips communicate
over the S2C or NFC-WI (wired
interface) bus.
The PN65K additionally includes SPI,
I
2
C, and UART interfaces. The read/
write range can be up to 50 mm with a
larger antenna and sufcient power. The
Smart Card Controller utilizes a Public
Key Infrastructure (PKI) co-processor
and a dual triple DES encryption key
coprocessor.
NXPs PN544 is pin-compatible with
the PN65N module but can work with
other secure chips. It also supports
the NFC-WI wired interface standard
ECMA373 for connecting external chips.
The NXP PN547, an improved version of
the PN544, features a longer read/write
48
range, a smaller footprint, and 50% less
power consumption.
Broadcoms BCM20791 and
BCM20792 are made with 40-nm
CMOS. These NFC controllers are
among the smallest (4 by 4 mm) devices
with extremely low power consump-
tion. Theyre designed to interface with
SIM cards or non-SIM security chips to
provide secure transmission. Theyre
also designed to pair with Broadcoms
BCM4330 Bluetooth, Wi-Fi, and FM
combo chip for handsets.
Texas Instruments offers a line of ICs
designed for NFC reader terminals. The
TRF7970A is an NFC/RFID transceiver
IC in a 5- by 5-mm, 32-pin quad at no-
lead (QFN) package. It complies with
all NFC stan-
dards including
NFCIP-1 (ISO/
IEC 18092) and
NFCIP-2 (IISO/
IEC 21481),
as well as
IDO14443A/B
and FeliCa.
Designed to
work with TIs
MSP430 micro-
controller or an
ARM MCU, it
has program-
mable output
power of 100
mW (20 dBm) and 200 mW (23 dBm). It
offers SPI and parallel interfaces and a
128-byte FIFO as well.
The austriamicrosystems AS3911 NFC
reader IC conforms to all NFC standards
including the EMVCo payment system
(Fig. 4). It features a capacitive sensor
and requires only 5 A to wake up in the
presence of a tag. The chip also has
fully automatic antenna tuning to opti-
mize performance. Its 1-W output power
eliminates the need for an external pow-
er amplier. And, it supports the VHBR
draft amendment to the 14443 standard,
allowing a data rate to 6.8 Mbits/s.
Based on the Microsoft Windows
Phone 7 platform, the Nokia Lumina 610
smart phone uses INSIDE Secures Mi-
NFC RF STANDARDS FOR CODING,
MODULATION, AND DATA RATE
NFC-Forum
standard
Polling or
listening
Coding Modulation Data rate
Carrier
frequency
NFC-A Polling Modified Miller ASK 100% 106 kbits/s 13.56 MHz
NFC-A Listening Manchester Load (ASK) 106 kbits/s
13.56 MHz
+ 848-kHz
subcarrier
NFC-B Polling NRZ-L ASK 10% 106 kbits/s 13.56 MHz
NFC-B Listening NRZ-L Load (BPSK) 106 kbits/s
13.56 MHz
+ 848-kHz
subcarrier
NFC-F Polling Manchester ASK 10%
212/424
kbits/s
13.56 MHz
NFC-F Listening Manchester Load (ASK)
212/424
kbits/s
13.56 MHz
(no subcarrier)
croRead v.3.4 NFC controller and Open
NFC protocol stack software (Fig. 5).
Also, INSIDE Secures SecuRead NFC
platform recently went through the EMV-
Co Platform Security Evaluation Process
and is now certied to work with 1.3 bil-
lion EMV-compliant cards and tags. It
comprises the SLE-97 embedded Se-
cure Element from Inneon, a Global-
Platform-compliant Java card operation
system coupled with the INSIDE Micro-
Read NFC controller and Open NFC
protocol stack.
Meanwhile, INSIDE Secures Micro-
Pass 4101-2K NFC tag complies with
the NFC Forums Type 4 tag require-
ments with 2 kbytes of memory. This is
enough to store long URLs, business
cards, phone numbers, or Wi-Fi or Blue-
tooth pairing information and other ap-
plication data that can be read by an
NFC-enabled device.
Finally, Marvells 88W8897 combina-
tion chip for smart phones includes
NFC in addition to its 802.11ac Wi-Fi as
well as Wi-Fi Miracast and location en-
gine (Fig. 6). Other vendors are adding
NFC their combo chips too.
Where Is NFC Headed?
The predictions for the use and growth
of NFC e-payments are very positive.
As the use of smart phones continues
to grow and as the incorporation of NFC
into these phones increases, the po-
tential for e-payment grows. Today, just
over 50% of the U.S. population has a
smart phone, but most do not have the
NFC component. Apples next iPhone is
expected to include NFC, though.
According to Jagdish Rebello of IHS
iSuppli, only 12% to 15% of current
smart phones feature NFC. He indicated
that there were 106.471 million devices
with NFC in 2011 and 232.507 million
projected for 2012. He estimates this will
grow to 989.142 million by 2016. More
phones with NFC will obviously boost
the e-payment movement.
Market research rm Gartner esti-
mates that the value of e-payments will
grow from $172 billion in 2012 to $600
billion in 2016. Gartner also expects the
number of global users to grow from just
over 200 million this year to more than
400 million in 2016.
6. Marvells Avastar 88W8897 incorporates
802.11ac 2x2 MIMO and beamforming to
produce upwards of 867-Mbit/s data streaming
for video. It includes additional Wi-Fi features in
addition to Bluetooth 4.0 and NFC radios.
In its new market reports, Juniper
Research projects total mobile payment
transactions to hit $1.3 trillion by 2017,
mostly thanks to physical goods sales.
Even then, however, mobile sales of
physical goods would still only account
for 4% of all global retail transactions
by 2017.
NFC tags and stickers may prove
to be very popular. Made of plastic
or paper, they include the NFC/RFID
chip that will provide some useful text,
graphic, ad, map, or URL. As more
smart phones get NFC, the use of
these stickers will grow. They are cheap
enough to be used on a wide range
of items or in strategic locations. The
sticker movement may be as big if not
bigger than the payment function.
There are several reasons why NFC
payment rollout is so slow, though. First,
theres a small number of NFC-enabled
smart phones. Eventually, though, most
smart phones will include an NFC radio.
Second, who will enable all those NFC
smart phones? Which payment service
will consumers select? The public still
needs to be educated about NFC.
Consumers then must select a service,
and selection alone is confusing when
so many options are available.
Furthermore, retailer point-of-
sale terminals must be upgraded to
include NFC readers. Its an expensive
conversion. Whos going to pay for it?
Finally, NFC e-payment must offer
some major benet. Consumers wont
have to carry as many credit cards,
and transactions wont take as long
(see Credit Cards Arent Obsolete
Yet at electronicdesign.com). But is
that enough to convince the masses
to switch to NFC payments? Many
consumers dont want to include their
nancial information on their phones,
even though NFC is just as safe as
their credit cards. In the end, traditional
credit cards wont disappear complete-
ly, but e-payments still will account for
many transactions.
For More Information
With more than 160 member compa-
nies, the NFC Forum (www.nfc-forum.
org) is a non-prot organization devoted
to promoting NFC, maintaining stan-
dards, and providing testing and certi-
cation for interoperability.
Sponsor members of the NFC Forum
include Barclaycard, Broadcom, INSIDE
Secure, Intel, MasterCard Worldwide,
NEC, Nokia, NXP, Renesas, Samsung
Electronics, Sony, STMicroelectronics,
and Visa. Principal members include
AT&T, HP, Inneon, Marvell, Qualcomm,
and Texas Instruments.
Whats Next in Payments, located at
PYMNTS.com, is devoted to news about
payment options including NFC. Q
A
utomotive electronic control ap-
plications require robust input
power conditioning. The circuitry
must absorb 100-V transients on
one hand, while providing a sta-
ble dc bus for a few hundred mil-
liseconds after the ignition is switched
off, so the CPU has the time to write
logged data into EEPROM before go-
ing off. The use of high-energy transient
absorbers usually meets the former re-
quirement, while an electro-mechanical
relay handles the latter.
The circuit described here offers novel
solutions for both requirements. It re-
places the costly and bulky transient
absorbers with an electronic switch and
the electromechanical relay with a sim-
ple RC timer that sustains the dc bus for
a preset amount of time after the ignition
is switched off.
The 12-V dc battery bus
enters the power-condi-
tioning stage through the
reverse-polarity protection
diode, D1 (see the gure).
Electronic switch Q2 is the
heart of the circuit block
responsible for bus-transient
protection. Transistors Q1
and Q3 form a logic block
that switches Q1 on and off
under different conditions.
Under normal conditions,
when the ignition switch is kept on, Q3
stays on and also maintains Q2 on. Q1
is usually off and turns on only when the
bus voltage exceeds 33 V dc, governed
by the 32-V zener, Z1 (under transient
and other overvoltage situations).
When the ignition key is turned off, Q3
stays on for a time governed by R5, C2,
and R6 to hold Q2 on. Since Q2 is di-
rectly connected to the battery through
D1, the circuit remains powered as long
as Q3 and, hence, Q2 is on.
However, the CPU senses the ignition
off condition through one of its inputs
and starts performing cleanup opera-
tions, such as storing the data logged
in RAM to EEPROM during the breather
provided to it by the R5-C2-R6 delay
circuitry. This clever scheme obviates
the need for the electromechanical relay
DESIGN IDEAS By Vishwas Vaidya
AUTOMOTIVE
POWER-CONDITIONING
CIRCUIT ELIMINATES
POWER-HOLD RELAY
thats usually used, reducing cost and
eliminating cumbersome wiring.
During an overvoltage/bus transient
condition exceeding 33 V dc, Q1 turns
on due to the 32-V dc zener in its base
circuit. It switches off Q2, protecting the
load-circuit from high voltages.
A tricky situation arises when a tran-
sient arrives while the CPU is logging
its data in RAM under normal opera-
tion. Q2 will switch off and power to
CPU will disappear, requiring a power-
on reset when the transient goes away
and Q2 restores the power. This is
highly undesirable since it would inter-
fere with the integrity of the data being
written in RAM.
This problem can be addressed by
putting the CPU to sleep during the
transient period (usually a few hundred
milliseconds) and waking it up when the
transient goes away. Doing this avoids a
hard reset under such situations.
Capacitor C8 keeps the CPU alive
during the sleep mode. An extremely
low CPU current avoids draining C8
during the transient period, avoiding
a CPU power-reset. The CPU can also
turn off output power drivers during the
transients to protect them from high cur-
rent spikes. Q
VISHWAS VAIDYA is an assistant gen-
eral manager of the electronics division at
Tata Motors Ltd. He is based in Pune, India.
He holds a masters degree in control engi-
neering from the Indian Institute of Technol-
ogy, New Delhi, India.
V
Batt
D2
D1
R5
R2
R1
R4
R6
R8
R7 Z3
C2
Z1
Gnd
Ignition
Q1
C1 C8
C7 C9
L1 L2
C13
R3
Z2
Q2
VR1
Logic supply
Load 1 Load n
Driver CPU
Q3
This power-conditioning circuit for an automotive electronic control unit replaces bulky transient
absorbers with an electronic switch, Q1, and an electromechanical relay with an RC timer circuit,
R5-C2-R6.
L
ow-power circuits commonly
use transformerless power sup-
plies. However, the use of earth
grounds in many of these circuits
creates a serious problem that is
often ignored.
The very popular Microchip Applica-
tion Note AN954 shows a circuit ground
(0 V), but the reference is silent about
whether it can be connected to the main
power supplys earth ground.
1
Similarly,
Apex Technologys Application Note 35
describes circuits that cannot be inter-
faced safely and legitimately with any
external test equipment.
2
The reasons for these problems are not
difcult to see. In the former (AN954), the
transformerless power supplys circuit
ground has high volt-
ages with reference to
neutral, and connecting
it to earth would cause
a dangerous situation.
Therefore, connecting
any external equipment
whose circuit ground
is connected to earth
would be downright
dangerous.
In the latter (Note
35), the transformerless
power supplys circuit
ground is connected
to neutral, so connecting any external
equipment to the circuit (such as an os-
cilloscope to troubleshoot the circuit)
would violate wiring regulations in most
countries and usually cause the circuit
breaker to trip if the external equipment
has its circuit ground connected to earth.
One solution might be to try to insert a
small resistor between the circuit ground
in question and the earth ground of the
external equipment. This prevents the
mains circuit breaker from tripping and
technically avoids violating wiring regu-
lations, but it causes a new problem
noise between the earth and neutral
lines. This noise gets into the circuit, de-
feating the purpose of the earth connec-
tion. Usually, the result is that the earth
SIMPLE STRATEGY
SAFELY CONNECTS
TRANSFORMERLESS-SUPPLY
CIRCUITS
By Ramkumar Ramaswamy
ground is, quite
simply, rendered
unusable.
A simple work-
around to this
problem assumes
that the transfor-
merless circuits
ground is tied to
neutral, which is
the much more
popular and pre-
ferred strategy for
transformerless
power supplies (see the gure). The
circuit is nothing fancy. Its a simple dif-
ferential amplier, but its use in this spe-
cic context is innovative. It illustrates
how a signal from equipment A can
safely and reliably be fed into equip-
ment B irrespective of which one has the
problematic circuit ground.
For example, A could be a signal gen-
erator feeding into a transformerless
circuit, B, in which case G1 is earth,
and G2 is neutral. If B is an oscilloscope
being used to troubleshoot a transfor-
merless circuit, A, G1 is neutral, and
G2 is earth. Either way, G1 is As circuit
ground and G2 is Bs circuit ground, as
well as the op-amp circuits ground.
The gain at both op-amp inputs is equal
and opposite, so the earth-neutral noise
(which is the noise at G1 with respect to
G2) cancels out completely and does
not appear at the input of B. Effectively,
the circuit allows for the safe use of both
neutral and earth terminals by making the
neutral-earth noise voltage appear as a
common-mode signal at the op-amp in-
put. This simple strategy overcomes the
two problems described above.
Any general-purpose, unity-gain sta-
ble JFET-input op amp with a low offset
voltage will work. Q
REFERENCES
1. Condit, Reston; Transformerless
Power Supplies: Resistive and Capaci-
tive, Application Note AN954, Micro-
chip Technology Inc., 2004.
2. AC-DC Power Supply Design, Appli-
cation Note 35, Apex Technology, Dec.
1999.
+
A
G1 G1
R1
100k
R3
100k
R2
100k
V+
V
R4
100k
G2 G2
G2
B
A simple op-amp circuit placed between a transformerless circuit with a
ground tied to neutral and a circuit with an earth ground will eliminate the
safety and noise problems associated with such connections.
APPLICATIONS By Roberto Scibilia and Lars Lotzenburger Texas Instruments
01.11.12 ELECTRONIC DESIGN EUROPE 56
N
arrow-band low-frequency pow-
er-line communications (NB LF
PLC) is a hot topic these days.
Engineers creating PLC modems
must consider which PLC stan-
dard (PRIME, G3, S-FSK) will be
used and which hardware will get the
best performance out of the design.
The modems power supply unit (PSU)
also is essential for good performance.
The PSU supplies the PLC modem as
well as all of the other components in
the device. In PLC-enabled smart me-
ters, for instance, the metrology part is
powered from the same supply. How-
ever, the modem plays a primary role in
the power supply design.
The rst parameter to check is the ef-
ciency of such a power supply over a
wide load range. Reception mode, when
the modem listens for the data packet at
the power line, requires about 200 mW.
However, the power consumption can
go up to almost 3 W if the modem needs
to transmit data packets into a power
line with low impedance (~2 ).
The modem is in listening mode most
of the time. It should have good efciency
at this working point. At the same time, it
must be able to respond very quickly to
more amps during data transmission.
The next parameter is the operating
frequency of the power supplys switch-
ing parts. Ideally, this frequency (and
the rst few harmonics) should not fall
into the frequency region where PLC
communication is performed. The NB LF
PLC frequency range is 0 to 500 kHz.
While most of the world uses the entire
band, in Europe the CENELEC commit-
tee has dened four bands going up
to only 148.5 kHz: Cen A (3 to 95 kHz),
Cen B (95 to 125 kHz), Cen C (125 to
140 kHz), and Cen D (140 to 148.5 kHz).
Cen A is reserved for metering operated
by utilities and their partners, while the
DESIGN
YOUR OWN
PLC-COMPATIBLE
PSU
remaining bands can be used for other
kinds of applications.
Knowing the devices end application
can help identify the right parts for the
PSU from the beginning. A design that
allows the selection of an operating fre-
quency by changing the values of pas-
sive components enables a common
PSU for different PLC applications. Of
course, working in the PLC transmission
band is allowed, but additional ltering
might be necessary.
Another important requirement is that
the mains lter of the PSU must not atten-
uate the PLC signal. By nature, the PLC
signal is injected at the same point the
PSU is connected to the mains. If a three-
phase smart meter design is planned,
the design must ensure the system works
even if one or two phases fail. A three-
phase PLC design will listen to all three
phases and drive (the same signal) into
all three phases at the same time.
Constant Switching Frequency
The design introduced here fullls
all of the requirements for a PLC-com-
patible PSU. We cannot use a typical
Green Mode controller, because it
adapts the switching frequency (FSW)
and the operating mode to the load.
For example, the UCC28600 works
at high line at 130 kHz for most of the
load conditions. But when the load is
very small, which is the case in receive
mode, the switching frequency decreas-
es to 40 kHz to stay in quasi-resonant
mode and prevent switching losses.
For very light loads, the frequency will
be clamped to 40 kHz and the control-
ler will enter a
burst mode.
The frequency
90 to 460 V
AC
50 to 60 Hz
Isolated AC-DC
with LM5021, 85%
F
SW
= 140 kHz
Regulation
16-V
winding
5-V
winding
Buck
TPS54040, 92%
Passive lter
LDO with
TPS76933
Buck
TPS62240,
91%
12 V 5%
500 mA
15 to 18 V
200 mA
(300 mApk)
80 mV p-p
3.3 V 3%
50 mA
(100 mA)
1.8 to 1.9 V
3%
150 mA
1. The proposed PSU design for PLC is
designed to maintain constant switching
frequency, even at light loads, to avoid
introducing harmonics to the PLC signal.
Rectier and lter #2
D1
MB8S
R7
10k
L2
1 mH
D4
MB8S
C2
0.1 F
1 kV
+
C3
47 F
350 V
+
C8
47 F
350 V
R19
10k
L5
1 mH
2. The 10-k
resistors damp
a second filter
after the rectifier
to prevent
unwanted
resonance in
the megahertz
range.
58
variation is even wider at low line, as the
controller sweeps the frequency back
and forth from 40 kHz to 130 kHz.
A better choice for this converter is
the LM5021, which can be congured
with constant switching frequency, with
or without pulse-skip mode and burst
mode. Burst mode keeps the switching
frequency constant inside each burst,
while the pulse skipping introduces a
frequency of FSW/n, where n = number
of skipped pulses.
Pulse skip mode should then be
avoided since it may generate sub-har-
monics in the PLC band. Figure 1 shows
the architecture of the whole PSU with
the description of all outputs needed by
the system.
An isolated yback converter provides
two outputs, 16 V and 5 V. The regulation
is connected on both outputs, but with
greater importance given to the 5-V out-
put. The 16-V output will supply the trans-
mitter section, which accepts 15 V to 18 V.
The 5-V section is down-stabilised to
3.3 V by means of a low dropout regula-
tor (LDO) and feeds the microcontroller.
A buck converter with a TPS62240 sup-
plies the digital part of the load.
The yback converter represents the
interface to the mains and the power line
communication. In typical ofine yback
converters, the input lter (Filter #1) has
classX capacitors connected directly
between line and neutral or line to line.
We want to avoid them, because they
create low impedance for the PLC sig-
nal and might reduce the signal-to-noise
ratio of the receiver.
An inductor should be the rst com-
ponent seen from the mains side, and
it must have higher reactance than the
PLC output impedance. A 10- fused
resistor is used to control the inrush cur-
rent and to protect the converter in case
of short circuit.
A second lter after the rectier, like
the rst one, has two functions: common
mode and differential mode, damped by
the 10-k resistors to avoid unwanted
resonance in the megahertz range (Fig.
2). The equivalent LC lter is achieved
by twice the inductance times the ca-
pacitor C2. The clamp network con-
nected to the MOSFETs drain is not the
typical RCD circuit, but a diode and tran-
sient voltage suppressor (TVS) in series.
E
f
c
i
e
n
c
y

(
%
)
1 2 3 4 5 6 7 8 9 10 11
Output power (W)
160 V dc no skip 160 V dc with skip 325 V dc no skip
325 V dc with skip 565 V dc no skip 565 V dc with skip
85%
80%
75%
70%
65%
60%
55%
3. The efficiency of the flyback converter varies
with and without skip-cycle mode. Disabling the
skip-cycle mode improves light load efficiency.
The advantage of this
setup compared to an
RCD network is that dur-
ing burst mode, an RCD
networks clamping ca-
pacitor would dissipate
energy into the parallel
resistor because it would
remain charged, or at
least, it wouldnt dis-
charge itself after each
burst cycle. In the diode
plus TVS case, the en-
ergy is lost during the
burst because there is
no storage element.
This strategy added to
the low startup current of
the LM5021. The burst mode feature lets
the yback stage gain efciency at light
load, even at high input voltages, which
is the case for a three-phase input.
By means of an NPN transistor and a
resistor, a small current is injected into
the current sense pin, sufcient to dis-
able the skip-cycle mode. If, in other
applications, the power supply was al-
lowed to take advantage of this working
mode, the resistor could be open and
the transistor unpopulated.
Figure 3 demonstrates efciency with
and without skip-cycle mode, showing
a clear improvement in the light load
efciency. Two groups of three curves
have been plotted by varying the input
voltage and the load. The
green lines show higher
efciency due to the skip-
cycle mode, while the
feature was disabled in
the red curves.
Figure 4 is a photo of
the prototype. It is impor-
tant to note that the lter
electrolytic capacitors are
quite large compared to
the available surface be-
cause the converter must
work from 90 V ac to 460
V ac, in single-phase,
three-phase, and line-to-
line delta connection. Q
ROBERTO SCIBILIA joined the Energy
System Labs at Italtel in 1990 after earn-
ing his masters degree in electronic en-
gineering, completed at the University of
Palermo in 1989. He is currently part of a
reference design team in Freising, Germa-
ny, as an application engineer, taking care
of off-line and isolated switching power
supply design.
LARS LOTZENBURGER graduated as a
diplom-engineer (FH) in 2001 and joined
Texas Instruments the same year as a sys-
tems engineer for data converter software.
In 2009 he internally changed his position
to become a systems engineer for power-
line communication (EMEA).
4. The filter electrolytic capacitors
in the prototype are quite large
compared to the available surface
because the converter must work
from 90 V ac to 460 V ac, in single-
phase, three-phase, and line-to-
line delta connection.
Publisher
bill baumann
00 1 212.204.4365
bill.baumann@penton.com
editor-in-Chief
Paul WhytoCk
00 44 20 8859 1206
paul.whytock@penton.com
editor-in-Chief, Electronic Design
Joe desPosito
00 1 212.204.4368
joe.desposito@penton.com
Creative direCtor
dimitrios bastas
00 1 212.204.4372
dimitrios.bastas@penton.com
managing editor
riChard gaWel
00 1 212.204.4381
richard.gawel@penton.com
assoCiate editor
sally Ward-foxton
Phone: +44 7979 553 549
sally.ward-foxton@penton.com
grouP ProduCtion manager
Justin marCiniak
00 1 913.967.1730
justin.marciniak@penton.com
ad ProduCtion manager
Julie gilPin
00 1 913.967.1373
julie.gilpin@penton.com
editorial offiCes
new york, ny 10011
00 1 212.204.4200
free subscription
status of subscription
address change
missing back issues
sales rePresentatives
e|design brand manager
traCy smith
Phone: 00 1 913.967.1324
tracy.smith@penton.com
united kingdom and european sales:
mark durham
Phone: +44 (0)7958 564137
mark.durham@penton.com
CirCulation Customer serviCe (live)
Phone: 00 1 847.763.9504
fax: 00 1 847.763.9673
electronicdesign@halldata.com
Archive
eleCtroniC design; maChine design;
miCroWaves & rf; sourCe esb; msd;
hydrauliCs & PneumatiCs; md&d; auto
eleCtroniCs; defense eleCtroniCs;
medi Cal desi gn; ee&t; di stri buti on
resourCes; global PurChasi ng;
eleCtroniC design China; engineering
tv; PoWer eleCtroniCs teChnology
Ad INdEX
digikey........................................................................................................................................................................................... 2
Micrel...........................................................................................................................................................................................11
Microchip..................................................................................................................................................................................... 5

Ede 1112

Transféré par

Informations du document

Description originale:

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Ede 1112

Transféré par

Droits d'auteur :

Formats disponibles

01 november 2012

Regul at or s s ave pr ec i ous

12 01.11.12 ElEctronic DEsign EuropE

Vous aimerez peut-être aussi