September 2011 / Volume 1, Issue 7
Wireless Technologies:
Mesh Or Mess?
Crystal Reassessment
The Multi-Stage Off Line LED Driver
Defence Electronics: Obsolescence Matters
september 2011
CONTENTS
4
Defence Electronics: Obsolescence Matters
8
Touch Technology: Detection by Deflection
14
Wiring Regulations : Get Wired For Wiring
What can be done to minimise the on-going problem of the impact of obsolescence in aerospace
and defence electronics?
Highlighting the key changes in new wiring regulations that come into effect on 1st January 2012.
Wireless: Wireless Technologies - Mesh Or Mess?
24
Crystal Maze: Crystal Reassessment
26
Driving LEDs: The Multi-Stage Off-Line LED Driver
34
Showtime: Electronics Production On The Up
Page 26.
The wireless scenario reviewed.
The time is now right for MEMS quartz crystals!
It’s relatively straightforward to design a basic LED
driver for a general lighting but more complex when
additional functionality is also required.
There’s a discernible upwards trend in electronics
production highlighted by the productronica buzz.
39
Grass Roots EMI: Meeting EMI For AC/DC Systems
42
Cabinets, Racks & Enclosures: Containing ESD
46
Cabinets, Racks & Enclosures: The Third Certainty
52
Page 20.
Combining a deformable panel/target layer and the
low power and simplicity of capacitive touch detection, creates a powerful solution.
20
50
Page 8.
Page 52.
Back to filter basics.
We look at coating finishes for moulded enclosures.
Thoughts on energy costs and how they fit into
Franklyn & Bullock’s definition!
Enclosures: Standard Boxes Clever For Simplicity
Making the case for modified standard over custom
enclosures.
Powering FPGSs: Powering Down FPGA Power
Supply Component Count
New µModule regulators are unleahed to reduce the
power supply component count for FPGA based
systems.
www.electronicspecifier.com
ElectronicSpecifier Ltd.
Ellion House, 6 Alexandra Road, Tonbridge, Kent TN9 2AA, UK
Editor: John Taylor BA, IEng, MIET
Email: editor@electronicspecifier.com
Tel/Fax: +44 (0) 1843 866 607
Copy Editor: Raj Joshi
Advertising: Tim Anstee
Email: tim@electronicspecifier.com
Tel: +44 (0) 1732 366 624 Fax: +44 (0) 1732 366 052
Publisher: Steve Regnier
Email: steve@electronicspecifier.com
Tel: +44 (0) 1732 366 617 Fax: +44 (0) 1732 366 052
Subscriptions
Subscribe to the Digital
Edition of ElectronicSpecifier
Distribution Blog
Find out what’s happening
in the world of distribution
electronicspecifierdesign | september 2011 | 3
defence electronics
Obsolescence Matters
O
Obsolescence management is an on-going problem within the aerospace and
defence electronics communities. What can be done to minimise its impact?
Sally Ward-Foxton investigates.
bsolescence is a much-talked about
issue in the aerospace and defence
electronics community. This problem
frequently manifests itself in this very specialised hi-rel market, since ensuring continuity of supply for electronic components is
considerably harder when the lifespan of
the system in question may be upwards of
40 years!
Over recent years there has been much consideration given to obsolescence management
which can undoubtedly save headaches further down the line. Good design, using devices with drop-in replacements from multiple
sources that can be used in the event that the
part goes EOL – end of life – are to be recommended wherever possible.
From the outset...
Working with manufacturers on obsolescence management from the outset can
bring huge benefits. Large semiconductor
manufacturers are setting up programmes
specifically for hi-rel customers to help them
service their equipment during its lifetime.
As an example, Maxim recently announced an obsolescence mitigation programme for military and aerospace OEMs
to provide longterm access to older technology. Maxim works with the OEM from the
time when the part is announced obsolete
to set aside wafers to support the obsolescence of any COTS plastic part.
This has to be done before the published
4 | september 2011 | electronicspecifierdesign
last time buy date but can ensure continuity of
supply for years to come.
Thinking out obsolecence
With a degree of forethought, it is
apparent that many of the obsolescence trials and tribulations can be
avoided altogether.
“Planning in advance for potential
obsolescence problems is the key to
ensuring their most effective management when they do arrive,” explains John Macmichael, Managing
Director of hi-rel specialist distributor
Solid State Supplies. “The most important thing is to develop an inhouse strategy for dealing with obsolescence and the starting point is to
analyse the factors that can affect
component sourcing.”
www.electronicspecifier.com
Macmichael suggests consideration of the so
called STEEPLE – social and demographic,
technological, economic, environmental, political, legal and ethical – factors that affect
the component supply chain. Some may seem
to be remote from the supply chain but the recent environmental disaster in Japan, for example, demonstrates how unpredictable the
world can be.
EOL planning
Proper EOL
event planning for
EOL can mitigate
the problems it
causes: every
semiconductor device will eventually reach the
EOL stage. However, the plain
fact is that obsolescence cannot
always be avoided. If a part has
already gone EOL
and you’ve missed the last time
buy date, what
are the options?
Firstly, always
check to see if the
manufacturer has
an alternative part
that would work in
your system. This
can save the cost
of having to evaluate a new supplier.
If a RoHS compliwww.electronicspecifier.com
ant version of the same product is available, it
may be possible to coat the terminations with
tin lead solder, for example, to eliminate the
problems caused by the development of tin
whiskers. Your distributor may be able to
arrange this for you.
If a screened part has been specified, it may
be possible to replace the obsolete component
with one that has been screened to a higher
level. As an example, an obsolete JAN
screened part may be replaced with a JANTX
or JANTXV screened alternative. A JANTX
screened part may be replaced with a JANTXV one. All this means is that the obsolete
part is being replaced with one that has been
screened to a higher level but is electrically
and mechanically identical.
“Of course, the higher the screening requirement, the more expensive the component, but
this may still be the most economical solution
to an obsolescence problem,” adds
Macmichael, cautioning that care must be
taken to understand the exact screening requirement because some end users may have
specific demands that are not addressed by
the general specifications.
Distribution sourcing
Asking an experienced distributor to look for
a source of replacement parts is also a viable
option. Reputable, manufacturer authorised,
franchised distributors are able to
provide full traceability for replacement parts.
“Distribution is a
small, close knit
community in the
UK and your preferred distrib- ☞
electronicspecifierdesign | september 2011 | 5
defence electronics
utor may be able to put you in touch with another reputable source for the components, or
even purchase them from that source on your
behalf,” Macmichael says.
Distributor credentials
Be sure to assess the distributor’s familiarity
with Mil-Spec and COTS parts and the related
standards that apply, particularly where screening is involved. Distributors that only dabble in
the hi-rel market are unlikely to have the depth
of experience and expertise to provide credible
advice regarding suitable parts that can be substituted for obsolete components.
It’s best to deal only
with distributors that
are subject to independent audits of their
processes and procedures. Those that are
qualified to BS EN ISO
9001:2008 and hold
AS9120 certification
will not risk their reputations by supplying
anything other than
components of the highest standard.
If all areas of the market have been scoured
and replacement parts cannot be found, there
are still options available. Companies can
partner with continuing manufacturers, who
can continue to manufacture hard to find and
EOL semiconductor devices with the authorisation of the original manufacturer.
“Companies like Force Technologies here in
the UK can hold the product in die form,”
Macmichael says. “They then complete the
product when it’s required, including carrying
out the required screening and test procedures. Once again, it’s very important only to
deal with reliable sources, that can provide
data enough to satisfy the required level of
6 | september 2011 | electronicspecifierdesign
confidence in the parts’ reliability. Force Technologies offers a range of options from a 57
point goods inward inspection to full parametric testing and screening to the various required standards.”
George Karalias, Director of Marketing &
Communications at Rochester Electronics explains, “Authorised continuing manufacturers
engage with the original semiconductor manufacturer to acquire the remaining inventory, including die, selected intellectual property, tooling, test programs and test equipment.”
Rochester has the largest inventory of wafers
and die in the world, and these are stored in
highly advanced,
climate controlled,
nitrogen purged
boxes. “The transfer
of technology ensures that semiconductor devices produced are exact
replicas of the original device in performance characteristics
and physical dimensions,” Karalias adds.
If wafer and die stocks have been exhausted, there is one final option.
Recreating the obsolete
Device recreation is a technique whereby
new semiconductor devices can be reverse engineered for the manufacture of a pin for pin
drop-in replacement with matched timing. Even
when the design archive is no longer available, engineers can redraw the device using
scope images, including scanning electron microscope imaging.
“The recreated part is guaranteed to deliver
the same specifications and performance as
the original,” Karalias maintains. “The end
product is a form, fit and function drop-in re-
www.electronicspecifier.com
placement for the original device. A side by
side comparison reveals the device is identical
to the original.”
Grey abandon...
A final word of warning: the so called grey
market is to be avoided at all costs! Counterfeiting is rife in electronics and counterfeit
components can be extremely difficult to spot.
Karalias points out that
electrical and reliability
testing is expensive, time
consuming and typically
requires assistance from
the original manufacturer, and that it’s impossible to test every device
that is delivered.
“The only sure way to
protect yourself, your
company brand, your
www.electronicspecifier.com
product’s reputation,
and your customers’
safety, is to buy only
from the original semiconductor manufacturer
or their authorised distributors,” he says.
With careful management from the outset, obsolescence problems can
be solved using one of
Above, George Kar- the solutions described
alias and left, John above, perhaps with
Macmichael.
help from a reputable
franchised distributor.
However, while component parts remain hard
to find, they will inevitably always attract
higher values and be susceptible to counterfeiting. It’s therefore extremely important to know
which are the sources that you can to trust. ❚❘
Return to contents page.
electronicspecifierdesign | september 2011 | 7
touch technology
Detection
by
Deflection
Metal Over Cap construction, an easy touch for Microchip.
I
The combination of a deformable panel/target layer and the low power and
simplicity of capacitive touch detection, create a powerful solution, says
Microchip’s Darren Wenn, for capacitive touch sensing with all-metal front panels.
n touch technology, combining a deformable panel/target layer with capacitive touch detection, eliminates the challenges associated with operating in the
presence of water and noise. Proximity triggering is replaced with actuation according
to a designer's exact specifications and the
system retains the low power operation of
capacitive touch.
Touch screen and touch panel interfaces
have leapt into the position of input medium of
choice in a raft of applications, driven to a
large extent by the success of portable products such as Apple's iPhone and iPad. Touch
driven interfaces, though, have a long history
8 | september 2011 | electronicspecifierdesign
prior to the current phase of widespread acceptance in consumer electronics: a wide variety of sensing technologies and a great deal
of ingenuity have been applied to the problem
over many years.
Explosive touch growth
Today's explosive growth in touch based
products has led to consumer product reviewers and even consumers pronouncing superiority for capacitive over resistive technologies.
Capacitive touch has many advantages. It is
low power, low cost, simple to implement, mechanically reliable and allows product designers more freedom and flexibility in matters
www.electronicspecifier.com
such as shape and placement of buttons, sliders, and the other control formats that are
now ubiquitous. In fact, there’s a plethora of
capacitive sensing approaches for designers
to choose from, all differing in the detail of the
techniques they employ, and all sufficiently different that their respective vendors can assert
patentable rights over them.
In all cases, however, the field effect nature
of capacitive touch sensing still imposes some
limitations, irrespective of the detailed implementation. Standard capacitive touch systems
do not normally work through metal and the
technology demands special software. In effect, signal processing of the sensor data is
used to maintain operation in environments
with radiated and/or conducted noise.
This is a far from trivial point when the touch
interface paradigm moves beyond the context
of the cool gadget into mission critical industrial control or medical applications.
Some other limitations of capacitive technologies include a limited ability to correctly detect
a touch in the presence of water or other contaminants or if gloves, especially heavy protective ones, are worn. And the fact that when
the location of a button is only delineated
graphically, a problem is created for visually
impaired users who rely on Braille.
Detecting panel flexure
A new technique that adapts proven detection circuitry with a simple yet novel physical
construction, overcomes all of these limitations
without compromising power consumption or
design simplicity. Taking the Metal Over Cap
approach, all of the circuitry is safely concealed behind a panel, usually, the front panel
of the product itself, yet it retains sensitivity to
touch and immunity to false input detection.
In a traditional capacitive system, the user
changes capacitance value by placing a finger
in close proximity to the sensor. The finger bewww.electronicspecifier.com
comes part of the sensing capacitor, either
forming the second plate of the capacitor or
changing the dielectric properties around an
existing double plate device.
The change in value is detected by a variety
of circuit techniques: for example, one elementary method is for the sense capacitor to set
the frequency of an oscillator, and the altered
capacitance shifts the frequency by a measurable amount.
In contrast to arrangements that use the touching finger directly as a capacitor plate, the
Metal Over Cap touch system uses a conductive target, suspended over the capacitive touch
sensors, to form the second plate of the capacitor. When a user applies pressure on the target
panel, the resultant deformation of the target
moves it closer to the capacitive sensor.
Practical design
A practical design requires a change in capacitor plate spacing and hence capacitance,
as capacitance varies inversely with distance
between plates of some 6% or more. The
change in capacitance is then measured by a
microcontroller. Figure 1A shows the basic
construction of a typical metal over capacitive
touch sensor, and Figure 1B depicts the deformation due to a user’s press.
Figure 1C shows an alternative configuration
that employs a metal target bonded to the
back of a plastic fascia layer. The target in this
configuration can be either a thin sheet of
metal bonded to the back of the plastic fascia
or a metal flashing or coating on the plastic
sheet. While the metal target still performs the
same electrical function as a metal fascia system, it is the physical characteristics of the
plastic which determine the mechanical deviation to the user’s press in this configuration.
The design of a successful Metal Over Cap
device and selecting appropriate materials
and dimensions, involves balancing a num- ☞
electronicspecifierdesign | september 2011 | 9
touch technology
ber of factors. The material from which the
panel is constructed
must be thin and flexible enough to deform
under a touch or light
press. However, the
configuration is sensi-
Figure 1A: Cross section of metal over capacitive - not touched.
Figure 1B: Cross section of metal over capacitive - touched.
hands may be gloved,
the touch might not always be centred on
the target.
Factors that affect
how the buttons interact include the elasticity of the target material and the adhesive’s adhesion to
tive to very small physical deformations, and
this requirement is
readily met.
The material should
be resilient and should
only deform within its
Figure 1C:
elastic limit, up to the
Cross section of metal over capacitive – touched using a plastic stylus.
anticipated maximum
applied pressure, so
that it returns to its flat condition after pressure bond the target to the spacer. If the target is
is removed.
too stiff and the adhesive is elastic, then a
force applied to button A will cause the target
Size & separation considerations
over sensor B to lift.
Other considerations include the size of the
buttons or other control devices and their sepaPractical constraint
ration, the adhesive system used to bond the
A practical constraint for panel design is that
spacer and target layer to the PCB and the thick- the spacing between buttons should be at least
ness of the spacer layer. Spacing between buta third to half the diameter of the buttons
tons is important in order to minimise movement
themselves and, with the very wide range of
on the sensors adjacent to the one selected.
materials that are usable for both panel and
Button size, layout and separation is imporspacer layers, designers should work closely
tant to minimise the panel deflection over the
with the manufacturer of the adhesive to select
sensors adjacent to the one that is pressed,
a suitable system.
bearing in mind that in an application where
Design of the spacer layer itself is a further ☞
10 | september 2011 | electronicspecifierdesign
www.electronicspecifier.com
LATEST
AND GREATEST
TECHNOLOGY
THE
WITH YOU TOMORROW.
At Farnell we celebrate the innovators; those of you who always look
to the future. With 100s of the newest technologies added each day
and next day delivery, you’ll be first to get your hands on the latest
components from the biggest brands.
www.farnell.com/new
www.element14.com
Design with the best
touch technology
Microchip: in touch with two capacitive systems.
consideration. The designer is attempting to
achieve flex in the target layer or panel, while
avoiding any movement in the PCB carrying the
fixed capacitor plates. As well as maintaining
the required gap, the spacer layer has to be
rigid: PCB material (FR4) or some other non-deformable plastic film can carry out this task.
Electrical design
Detection of the shift in capacitance can employ a number of techniques similar to those
used in other capacitive touch sensor architectures. Microchip has two systems and both
work well using Metal Over Cap construction.
In the capacitive voltage divider approach,
sensor capacitance is connected in parallel
with an analogue/digital converter's sample
and hold capacitor. The change in simply
touched versus untouched capacitance, ap-
12 | september 2011 | electronicspecifierdesign
pearing as a voltage shift at the ADC's input
and is converted to a digital value. The charge
time measurement unit approach uses a constant current source and measures the capacitance via its charge/discharge curve. In both
cases, software is readily available to incorporate into the microcontroller’s coding.
Both detection techniques yield a shift of
60~100 counts in a 12~14 bit result, for a 6%
capacitance change. This is sufficient to provide well discriminated touch detection, although care is needed in design of the system
to reduce mechanical and electrical noise.
AC and dc grounding of the target layer,
sensing PCB screening, careful connection
routing and good practice in selecting bypass
❚❘
capacitors will deliver a robust design.
More from Microchip Technology
Return to contents page.
www.electronicspecifier.com
wiring regulations
Get Wired For Wiring...
Geoff Cronshaw, the IET’s Chief Electrical Engineer, highlights the key changes in new wiring regulations that come
into effect on 1st January 2012. And they are highly relevant
to anyone working within electrical installation.
T
he recently launched 17th edition of the
IET’s Wiring Regulations, BS 7671:2008,
incorporate Amendment No 1: 2011.
And these new wiring regulations, jointly published with the BSI, set a new standard for all
new electrical installations in the UK.
The amended Wiring Regulations, BS
7671:2008 (2011) are
based on CENELEC Harmonised Documents, HD,
and IEC standards for Europe and the international community.
The standardisation
process is perpetual and
thus standards are continually evolving requiring
national standards bodies
to incorporate the technical changes.
Therefore, these new
amended Wiring Regulations embody the most recent changes made at a
European and international level.
One of the major changes that those working within the electrical installation industry
should be aware of is Section
534. This section sets out
the requirements for the
control of transient
over-voltages which
14 |
can affect all modern electronic components
and processors.
This new section gives detailed requirements
on the selection, erection and co-ordination of
surge protective devices, SDPs.
Section 534: over-voltage protection devices
This new section in the
regulations has been implemented by CENELEC
in HD 60364-5-53. An
SPD is intended to limit
transient over-voltages
and divert damaging
surge current away from
sensitive equipment.
With equipment now
more precise and with
little tolerance on the
components, and hence
less robust than it ever
was, SPDs are more important than ever.
SPDs must have the
necessary capability to
deal with the surge current levels and durations expected at their
point of installation. They can operate in one
of two ways, based on the component technologies within the devices.
❏ One way is as a voltage switching device.
Under normal conditions, the device is ☞
september 2011 | electronicspecifierdesign
www.electronicspecifier.com
Circuit Note
CN-0209
Circuits from the Lab™ reference
circuits are engineered and tested for
quick and easy system integration
to help solve today’s analog, mixedsignal, and RF design challenges. For
more information and/or support,
visit www.analog.com/CN0209.
Devices Connected/Referenced
ADP1720
High Voltage, Low Dropout Linear Regulator ADG442
LC2MOS Quad SPST Switches
AD8275
G = 0.2, Level Translation ADC Driver
AD8676
Dual, Rail-to-Rail Output Op Amp
REF194
Precision, 4.5 V Voltage Reference
AD7193
24-Bit ∑-∆ ADC
AD8617
ADT7310
Low Noise, Rail-to-Rail Input/Output Amp
16-Bit Digital Temperature Sensor
ADuM1400 Quad-Channel Digital Isolator
ADuM1401 Quad-Channel Digital Isolator
ADG1414
Serially-Controlled Octal SPST Switches
Fully Programmable Universal Analog Front End
for Process Control Applications
EVALUATION AND DESIGN SUPPORT
CIRCUIT DESCRIPTION
Circuit Evaluation Boards
CN-0209 Circuit Evaluation Board (EVAL-CN0209-SDPZ)
System Demonstration Platform (EVAL-SDP-CB1Z)
This circuit provides a fully programmable universal analog front
end (AFE) for process control applications supporting 2-, 3-,
and 4-wire RTD configurations, thermocouple inputs with cold
junction compensation, unipolar and bipolar input voltages, and
4 mA-to-20 mA inputs as shown in the configuration diagram of
Figure 2. The ADG1414, a serially controlled octal SPST switch, is
used to configure the selected measurement mode.
Design and Integration Files
Schematics, Layout Files, Bill of Materials (zip file)
CIRCUIT FUNCTION AND BENEFITS
The circuit shown in Figure 1 (shown on page 2) provides a fully
programmable universal analog front end (AFE) for process control
applications. The following inputs are supported: 2-, 3-, and
4-wire RTD configurations, thermocouple inputs with cold
junction compensation, unipolar and bipolar input voltages, and
4 mA-to-20 mA inputs.
Today, many analog input modules use wire links (jumpers) to
configure the customer input requirements. This requires time,
knowledge, and manual intervention to configure and reconfigure
the input. This circuit provides a software controllable switch
to configure the modes along with a constant current source to
excite the RTD. The circuit is also reconfigurable to set commonmode voltages for the thermocouple configuration. A differential
amplifier is used to condition the analog input voltage range to
the ∑-∆ ADC. The circuit provides industry-leading performance
and cost.
Because of the voltage gain provided by the AD8676 and AD8275,
the design is particularly suitable for small signal inputs, all types
of RTDs, or thermocouples.
The AD7193 is a 24-bit ∑-∆ ADC that can be configured to have
four differential inputs or eight pseudo differential inputs. The
ADuM1400 and ADuM1401 provide all the necessary signal
isolation between the microcontroller and the ADC.
The circuit also includes standard external protection and
is compliant with the IEC 61000 specifications.
Voltage Measurement
This circuit supports the measurement of unipolar and bipolar
signal ranges up to ±10 V. The input signal goes through a signal
conditioning stage before conversion by the AD7193 ADC. The
AD8676 amplifier buffers the inputs before the gain stage. The
AD8275 is used to level shift the input signal and provides gain
so that it matches the input range of the AD7193. The AD8275
output is biased with a common-mode voltage connected to its
REF pin. This voltage is generated by the REF194 precision
4.5 V reference.
RTD Measurement
As shown in the connection table, this circuit supports 2-, 3-,
and 4-wire RTD configurations. In this case, the transducer is
a 1000 Ω platinum (Pt) RTD (resistive temperature device). The
most accurate arrangement is a 4-lead RTD configuration. In the
application shown, an external 200 µA current source provides
the excitation current for the RTD, and the AD7193 is operated
at gain of 16 to maximize the dynamic range in the circuit. The
AD8617 amplifier is configured as a current source when the RTD
measurement mode is selected. It is reconfigured in closedloop to set the common-mode voltage when the thermocouple
measurement is selected. The AD8617 is a dual low noise
amplifier so that it can drive both input channels available on
the board. The resistor configuring the current source must
have a low temperature coefficient to avoid drift errors in the
measurement circuit.
Circuits from the Lab™ circuits from Analog Devices have been designed and built by Analog Devices engineers. Standard engineering practices have been employed in the design and construction of each
circuit, and their function and performance have been tested and verified in a lab environment at room temperature. However, you are solely responsible for testing the circuit and determining its suitability
and applicability for your use and application. Accordingly, in no event shall Analog Devices be liable for direct, indirect, special, incidental, consequential or punitive damages due to any cause whatsoever
connected to the use of any Circuits from the Lab circuits. (Continued on last circuit note page)
www.analog.com
Circuit Note
CN-0209
Figure 1. Universal Programmable Analog Front End for Process Control
Applications (Simplified Schematic: All Connections and Decoupling Not Shown)
To view figures 1, 2 and 3 at a larger size click here.
Thermocouple Measurement
In a thermocouple application, the voltage generated by the
thermocouple is measured with respect to an absolute reference,
provided externally to the ADC. The cold junction compensation is
implemented using the ADT7310 16-bit temperature sensor. Because
the signal from the thermocouple is small, and to maximize the
dynamic range in the circuit, the AD7193 is operated at its highest
gain range of 128. Because the input channel is buffered, large
decoupling capacitors can be placed on the front end, if required, to
eliminate any noise pickup that may be present in the thermocouple
leads. The common-mode voltage for the thermocouple
measurement is provided by the AD8617 amplifier.
4 mA-to-20 mA Current Measurement
This circuit also supports 4 mA-to-20 mA current measurement. The
current is converted to a voltage using an on-board sense resistor. To
use the full dynamic range of the ADC in the current measurement
mode, a 200 Ω resistor is used.
The sense resistor must have a low temperature coefficient to avoid
temperature drift errors in the measurement circuit.
Figure 2. Analog Input Configuration Table
Regulator and Reference Selection
The ADP1720 was chosen as the 5 V regulator for this circuit.
The ADP1720 is a high voltage micropower linear regulator
particularly suitable for industrial application.
The 4.5 V REF194 was chosen as the reference for this circuit,
and the E-grade device has an initial accuracy ±2 mV at 25°C
and a drift of 5 ppm/°C maximum. It is a low dropout device and
consumes less than 45 μA, with performance specifications over
−40°C to +125°C.
Isolation
The ADuM1400 and the ADuM1401 are quad-channel digital
isolators based on Analog Devices’ iCoupler® technology. These
are used to provide isolation between the field side and the
system microcontroller, with an isolation rating of 2.5 kV rms.
Four wires are used through the ADuM1400, all for transmit
(SCLK, DIN, ADG1414, ADT7310). Four wires are used through the
ADuM1401: one for transmit (AD7193) and three for receive (INT1,
INT2, DOUT). The DIN, DOUT, and SCLK lines are connected to the
SPORT interface.
Table 1. Measured Performance Based on 1000 Samples
Input
±10V Input Range
4 mA to 20 mA Range
RTD
Thermocouple
AD7193 Configuration
Gain = 1; 50 Hz and 60 Hz rejection; output data rate = 50 Hz
Gain = 1; 50 Hz and 60 Hz rejection; output data rate = 2.63 Hz
Gain = 16; 50 Hz and 60 Hz rejection; output data rate = 2.63 Hz
Gain = 128; 50 Hz and 60 Hz rejection; output data rate = 2.63 Hz
RMS Noise (nV)
7940
931
243
220
Figure 3 shows a histogram plot of the AD7193 output performance when configured in bipolar input
mode with the inputs connected to ground. This histogram shows the effects of input-referred noise. The
effective resolution achieved in this mode is 19.2 bits.
Table 1 shows the performance for other operating modes based on 1000 data samples from the ADC.
This design also includes external protection such as standard protection diodes and transient voltage
suppressors (TVS devices) to enhance the robustness of the circuit. Refer to the schematics and other
resources in the CN0209 Design Support Package: www.analog.com/CN0209-DesignSupport (zip file)
Figure 3. Noise Distribution Histogram,1000 Samples AD7193 50 Hz Data Rate, Gain = 1, Input = 4.5 V Reference
Effective Resolution (Bits)
19.15
22.24
20.29
19.23
Circuit Note
CN-0209
CIRCUIT EVALUATION AND TEST
This circuit uses the EVAL-CN0209-SDPZ circuit board and the
EVAL-SDP-CB1Z System Demonstration Platform (SDP) evaluation
board. The two boards have 120-pin mating connectors, allowing
for the quick setup and evaluation of the circuit’s performance. The
EVAL-CN0209-SDPZ board contains the circuit to be evaluated,
as described in this note, and the SDP evaluation board is used
with the CN0209 evaluation software to capture the data from the
EVAL-CN0209-SDPZ circuit board.
Equipment Needed
• PC with a USB port and Windows® XP or Windows Vista®
(32-bit) or Windows® 7 (32-bit)
• EVAL-CN0209-SDPZ Circuit Evaluation Board
• EVAL-SDP-CB1Z SDP Evaluation Board
• CN0209 Evaluation Software
• Power supply: +15 V and –15 V
• RTD temperature sensor
• Thermocouple
Getting Started
Load the evaluation software by placing the CN0209 Evaluation
Software disc in the CD drive of the PC. Using “My Computer,”
locate the drive that contains the evaluation software disc and
open the Readme file. Follow the instructions contained in the
Readme file for installing and using the evaluation software.
Functional Block Diagram
See Figure 1 of this circuit note for the circuit block diagram,
and the file “EVAL-CN0209-SDPZ-SCH-Rev0.pdf” for the circuit
schematics. This file is contained in the CN0209 Design Support
Package: www.analog.com/CN0209-DesignSupport (zip file)
Setup
Connect the 120-pin connector on the EVAL-CN0209-SDPZ circuit
board to the connector marked “CON A” on the EVAL-SDP-CB1Z
evaluation (SDP) board. Nylon hardware should be used to firmly
secure the two boards, using the holes provided at the ends of the
120-pin connectors. With power to the supply off, connect a +15 V
power supply to the pin marked “+15 V,” a −15V power supply to
the pin marked “−15 V”and “GND” on the board. Connect the USB
cable supplied with the SDP board to the USB port on the PC. Note:
Do not connect the USB cable to the mini USB connector on the
SDP board at this time.
Test
Apply power to the ±15 V supply connected to the EVAL-CN0209SDPZ circuit board. Launch the evaluation software and connect
the USB cable from the PC to the USB mini-connector on the
SDP board.
Once USB communications are established, the SDP board can
now be used to send, receive, and capture serial data from the
EVAL-CN0209-SDPZ board.
Voltage Measurement
If you want to measure the noise of the voltage measurement
circuit, connect both inputs J3 and J4 to the ground. Then, click on
the button of the matching channel of the software: either V1 (if you
are using channel 1) or V2 (if you are using channel 2).
If you want to measure a voltage, connect both inputs J3 and J4 as
shown in Figure 2, the analog input configuration table. Then, click
on the matching button of the software as previously explained.
Results are displayed as a waveform and a histogram. You have the
option to select the scale of the voltage result between µV, mV, and
V using the switching button.
RTD Measurement
If you want to measure the temperature through an RTD
temperature sensor, connect inputs J1, J2, J3, and J4 as shown in
Figure 2. There are three different configurations of connection as
you are using RTD 2-, 3-, or 4-wire. Then, click on the matching
button of the software (RTD1 for channel 1, RTD2 for channel 2).
The switching button above the waveform allows you to display the
result in Fahrenheit, Celsius, or Kelvin.
Thermocouple Measurement
If you want to measure the temperature through a thermocouple,
connect inputs J1, J2, J3, and J4 as shown in Figure 2. Select the
type of thermocouple you are using (B, E, J, K, R, S, T, N). Then,
click on the TC button of the software (TC1 for channel 1, TC 2 for
channel 2).
The switching button above the waveform allows you to display the
result in Fahrenheit, Celsius, or Kelvin.
Current Measurement
If you want to measure a current, connect both inputs J5 and J6,
as shown in Figure 2. Then, click on the matching button (I1 for
channel 1, I2 for channel 2). You can select the scale of the current
result between µA, mA, and A using the switching button.
Information regarding the SDP board can be found in the
SDP User Guide.
Circuits from the Lab circuits are intended only for use with Analog Devices products and are the intellectual property of Analog Devices or its licensors. While you may use the Circuits from the Lab circuits
in the design of your product, no other license is granted by implication or otherwise under any patents or other intellectual property by application or use of the Circuits from the Lab circuits. Information
furnished by Analog Devices is believed to be accurate and reliable. However, Circuits from the Lab circuits are supplied “as is” and without warranties of any kind, express, implied, or statutory including, but
not limited to, any implied warranty of merchantability, noninfringement or fitness for a particular purpose and no responsibility is assumed by Analog Devices for their use, nor for any infringements of patents
or other rights of third parties that may result from their use. Analog Devices reserves the right to change any Circuits from the Lab circuits at any time without notice but is under no obligation to do so.
©2011 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners.
www.analog.com
wiring regulations
open circuit but, at a certain threshold voltage, the SPD conducts and diverts the current
through it to earth. It has two states, On and
Off, and hence the name voltage switching.
❏ Another way is as a voltage limiting device.
Voltage limiting SPDs again present an open
circuit under normal conditions. When an overvoltage is detected, the device begins to conduct, dropping its resistance dramatically such
that the over-voltage is limited and the surge
current is diverted away from the protected
equipment.
Section 534 details the selection requirements for SPDs to ensure that the correct type
of device is installed at the correct position
within an installation. The section also stipulates a number of requirements for the connection of SPDs, depending on the type of supply
and earthing arrangements and for installation
in circuits with RCDs.
This section also sets out the requirements for
the connection of SPDs: one example being
the requirements for the critical length of connecting conductors.
To gain maximum protection, it is stipulated
that the supply connecting conductors to the
SPD shall be kept as short as possible, to minimise additive inductive voltage drops across
conductors.
Regulation 534.2.10 sets out the requirements for the cross-sectional area of connecting conductors. The connecting conductors of
SPDs shall either:
i) have a cross-sectional area of not less than
4mm² copper, or equivalent, if the cross-sectional area of the line conductors is ≥ 4mm², or
ii) have a cross-sectional area not less
than that of the line
18 |
conductors, where the line conductors have a
cross-sectional area of <4mm².
Section 444: measures against
electro-magnetic disturbances
This entirely new section in the regulations
was brought about by developments in CENELEC - HD 60364-4444 Measures
Against Electro-magnetic Disturbances. In
1996, the EMC Directive (89/336/EEC)
was issued requiring
electromagnetic compatibility of products
being sold across Europe. The Directive relies
on harmonised standards that products must
comply with.
The CE Mark, for example, is a direct result
september 2011 | electronicspecifierdesign
www.electronicspecifier.com
of the work undertaken to remove trade barriers and, hence, allowed development of the
EMC Directive.
In 2004, the EMC Directive was modified as
2004/108/EC and came into force in July
2007; the main change was the inclusion of
fixed installations.
Subsequently, Section 444 of the Wiring
Regulations has been developed to provide
the basic requirements and recommendations to enable the avoidance and reduction of electro-magnetic disturbances and
meet the Directive.
The section includes guidance for the location and sources of electro-magnetic disturbances relative to the positioning of other
equipment as well as measures for minimising
www.electronicspecifier.com
electro-magnetic disturbances.
Section 729: operating &
maintenance gangways
Section 729 is intended to ensure a safe
working environment for skilled or instructed
persons working in areas such as switch rooms
and control gear assemblies where, historically, not much space has been allocated.
The Section has been incorporated into BS
7671:2008(2011) to align with the European
Harmonised Document.
Regulation 15 of the Electricity at Work Regulations 1989 requires that adequate working
space, means of access and lighting shall be
provided at all electrical equipment on which,
or near which, work is being done.
This section sets out minimum requirements
for the provision of adequate working space,
means of operational and emergency access
and evacuation at, or near, electrical equipment which is being worked on.
The amended regulations outline new requirements for the width of gangways and access areas which are to be adequate for work,
operational access, emergency access and
egress and for the movement of equipment.
For further information
or help about the IET
Wiring Regulations, 17th
Edition, BS 7671:2008 incorporating Amendment
No 1: 2011, call the free
Technical Helpline on
01438 765 599 or email
technical@theiet.org. The
new regulations are also
available for purchase at
www.theiet.org/amend1.
Guidance Notes and training courses are also
available through the IET: more information at
❚❘
www.theiet.org/electrical.
Return to contents page.
electronicspecifierdesign | september 2011 | 19
wireless
Wireless Technologies: Mesh Or Mess?
O
John Sharples reviews the wireless scenario.
ver the past five years it has not been
hard to notice the impact ‘wireless’
has had on all of our lives. We touch
wireless every day. Indeed it is difficult to find
a place on Earth not covered by a signal, a
place without access to wireless technology.
Radio, TV, satellites, GPS, GSM, Bluetooth,
Wi-Fi, RFID and keyless entry are just a few examples of wireless technologies we come into
contact with on the daily grind.
ISM networks
A phenomenon over recent years is a trend to
discover solutions using low, or indeed less
power wireless networks working in ISM bands. Smart
houses, smart buildings, smart
meters, smart grids, smart
anything, they all need wireless inside to be smart, or at
least to be qualified for the
smart moniker.
So, do we need
ISM wireless networks? The answer
to that, of course, is
yes, we do. To feel
more comfortable, to
be more flexible, to
take control of existing building infrastructure without disruption and dust!
Add to this a need to collect data from remote
sensors which could enable a household to save
money by better control of home equipment,
heating, lighting and security it is easy to see
20 | september 2011 | electronicspecifierdesign
how wireless creates a path towards convenient,
cost-effective living.
Low power wireless?
Environmental concern, cost and battery life
are driving future development of low
power technologies.
In common with
many other industries, wireless technologies must reduce
their power consumption. But what
does ‘low power
wireless’ mean?
Take a closer look at the marketing brochures of wireless technology vendors. They redefine existing standards for low power
wireless using phrases such as
ultra-low, extra-low or super-low.
Sometimes ultra-low power fits in
mA to low power, sometimes 100 times higher.
Wireless devices are often in receive mode
for many real applications and transmit data
only a few times a day. Therefore, the key to
low power wireless from a technical point of
view is low consumption in the receive mode,
best stated in some non-marketing units, in the
order of µA.
Mesh, mesh, mesh…
Until recently only a select few scientists used the
word ‘mesh’ in connection with wireless. However,
now every trained electronic component sales person claims to be an expert in selling products ☞
www.electronicspecifier.com
wireless
supporting wireless mesh networks, WMN.
Every technically educated person is able to
understand WMNs and there are a lot of mesh
supporting technologies gaining attention, especially from purchasing managers.
Wireless mesh
networks
Currently, many
different wireless
technologies available in the market
declare WMN support. So, it seems
that commercial
availability is quite
wide and should
satisfy every technical need. On the other
hand, by going deeper
into specifications and
datasheets, an engineer would find certain
limitations.
Some technologies support only a few – 2~6 – routing hops. Some declare an infinite number of
routing hops and some don’t declare anything
and avoid broaching the subject.
Wireless mesh networks & physics
The most popular ISM band for WMN seems
to be 2.4GHz. Its worldwide availability is attractive and promises many benefits. Achievability of relatively high transmitting power seems to
be advantageous, and it is a nice tool for engineers looking for simple solutions.
However, Messrs Lorentz, Faraday, Maxwell,
Hertz, Einstein and others, all defined the physical behaviour of our world and discovered
some interrelations and limitations: for example,
wavelength and the ability of radio wave to
penetrate walls.
22 | september 2011 | electronicspecifierdesign
Our world is full of walls, which is why
2.4GHz and the marketing driven enthusiasm
for it has slightly decreased. And new ISM
sub-GHz bands were re-discovered shortly
after the first low power WMNs had been
practically tested!
The best wireless technology
Claiming something to be best is always dangerous! In technical forums, this usually results in
heated discussion. Supporting this claim would
always appear difficult, realistically it is not, in
fact it is relatively easy.
The best wireless technology should be read
as the most convenient wireless technology for a
specific purpose or target application. No
(wireless) technology is universally convenient for every purpose!
Wireless mesh networks and the future
Although there is no best universal
wireless technology,
there will be many
co-existing in the future. And current
ISM bands will become crowded with
the expected boom
of WMN.
If we are at the
very beginning of
that expected
boom, and are already experiencing
interferences, what can we expect when the
boom of wireless mesh networks really comes to
a head? Is it advantageous to use transmitting
power on the same edge as thousands of future
❚❘
wireless neighbours?
John Sharples is MD of Low Power Radio Solutions.
More from LPRS
Return to contents page.
www.electronicspecifier.com
IR’s Solution for Energy Saving Drive
Automotive DirectFET®2 Power MOSFETs
Part Number
VDS
RDS(on) Max
@10VGS
ID max.
@TC = 25°C
Qg typ.
@10VGS
Package
AUIRF7669L2
100 V
4.4 m1
114 A
81 nC
DirectFET L
AUIRF7759L2
75 V
2.3 m1
160 A
200 nC
DirectFET L
AUIRF7739L2
40 V
1 m1
270 A
220 nC
DirectFET L
AUIRF7736M2
40 V
3.1 m1
141 A
83 nC
DirectFET M
600V High Voltage IC for Switching Stage Drivers
Part Number
Output
Current
Description
VCC UVLO
Package
AUIRS2191S
High Speed High and Low Side
+3.5 / -3.5 A
8.2 V
SOIC16N
AUIRS21811S
High Speed High and Low Side
+1.9 / -2.3 A
8.2 V
SOIC8
600V Automotive IGBTs for Switching Stage
Part Number
IC @TC=100°C
VCE(on) typ.
Package
AUIRGP35B60PD
34 A
1.85 V
TO-247
AUIRGP50B60PD1
45 A
2.00 V
TO-247
25V Low Voltage IC for Switching Stage Drivers
Part Number
Description
Output
Current
Package
AUIRS4426S
Dual Channel Low Side
+2.3 / -3.3A
SOIC8
AUIRS4427S
Dual Channel Low Side
+2.3 / -3.3A
SOIC8
AUIRS4428S
Dual Channel Low Side
+2.3 / -3.3A
SOIC8
For more information call +49 (0) 6102 884 311
or visit us at www.irf.com
Features
t"VUPNPUJWF2RVBMJGJDBUJPOT
t)7*$XJUIJOUFHSBUFEQSPUFDUJPOT
t0QUJNJ[FEDPQBDL*(#5GPSIJHI
GSFRVFODZ4.14PQFSBUJPO
t&YUSFNFMZMPX3%40/
'&5BOEIJHI
DVSSFOUSBUJOHTGPSTZODISPOPVT
SFDUJGJDBUJPOTUBHF
THE POWER MANAGEMENT LEADER
crystal maze
M
Crystal Reassessment
The time is right for MEMS quartz crystals, says David Potts, Divisional
Marketing Manager of Anglia Components.
EMS manufacturing has enabled the
production of high performance and
ultra-miniature timing components in small surface mount packages. The higher cost of these
components, however, has encouraged designers to continue to use the old
favourite HC-49 crystals and these are are
often the highest component on the board.
Looking at total cost of
ownership, crystals such
as Epson Toyocom’s
QMEMS range is certainly no more costly
than more traditional
products. These devices
also deliver enhanced
performance and stability into the bargain. Thus
they are seeing increased adoption
around the world, although the UK is somewhat
slower in following the trend. Perhaps now is the
time to reassess the choice of timing component.
MEMS moves to quartz
Most commonly, micro-electro-mechanical or
MEMS systems are based on silicon, polymers or
metals. Epson Toyocom pioneered the application of these manufacturing techniques to quartz
crystal and launched its first QMEMS product in
October 2006.
Quartz is hard and lends itself well to precision
24 | september 2011 | electronicspecifierdesign
manufacturing. It is also highly stable in response
to changes in temperature, shock, vibration
and exposure to chemicals.
With these characteristics, the application of MEMS technology has made
substantial miniaturisation possible.
MEMS semiconductor manufacturing processes are based
on wafer batch production, allowing productivity to be maximised and
many different variants
to be produced using
the same tools. As size
reduces, output can be
increased by increasing
the number of individual
components per wafer.
These tools can make
much smaller quartz devices than is possible
Ceramic Crystal
using conventional meStructure.
chanical machining.
Such devices have
proved extremely popular in handheld consumer
electronics, due to their exceptionally small size
and low power consumption – and they are
growing in popularity in electronics applications
of all kinds.
Precision fabrication improves stability and
consistency. For example, photo etching technology enables a 20% reduction in the size of
tuning fork crystals.
Machine processed tuning fork crystal designs
are two dimensional in construction and this
means that the smaller a tuning fork, the greater
www.electronicspecifier.com
the impedance, and this adversely affects performance. Applying 3D MEMS technology can
increase the overall surface area and helps hold
down the increase in impedance.
QMEMS benefits
As a result, the benefits of QMEMS extend beyond the production phase into the service life of
the product. Due to the consistent and very high
precision manufacturing approach, QMEMS
crystals are extremely reliable and stable over
time. Equivalent series resistance, ESR, for example, is stable even at very low drive levels, and is
unaffected by exposure to solder reflow operations or ageing.
SMD products are designed to endure the
rigours of reflow soldering, and will withstand exposure to temperatures of over
150°C, providing, that is, that
the recommended reflow
profile is adhered to.
Notably, the
development of
QMEMS devices
has led to Epson Toyocom capturing some 24%
of the global market but its share of the UK market is lower. This is probably because industrial
electronic designs dominate the market in this
country rather than the portable consumer electronics applications that dominate the volume, if
not the value, of the world electronics market.
So using MEMS based quartz crystal timing
technology provides a very competitive total
cost of ownership, as well as better stability
and performance than most alternatives. Now
it is time that UK designers reassessed the
❚❘
value of this technology!
More from Anglia Components
Return to contents page.
www.electronicspecifier.com
electronicspecifierdesign | september 2011 | 25
driving leds
The Multi-Stage
Off-Line LED
Driver
It’s relatively straightforward
to design a basic LED driver for
general lighting but it’s far
more complex when additional
functionality such as phase cut
dimming and power factor
correction are also required.
Peter B Green explains.
A
non-dimming LED driver without
power factor correction generally
comprises an off-line switching power
supply configured to regulate the output at a
constant current. It’s not much different from
a standard off-line switching power supply
such as those commonly used in ac/dc adaptors. And such designs can be based on standard SMPS circuit topologies like the buck,
boost or flyback converter.
On December 3rd 2009 the US Department
of Energy released the final version of Energy
Star Program Requirements for Integral LED
Lamps. This mandated that power factor must
be better than 0.7 for domestic applications for
LED drivers. The requirement for industrial applications is expected to be better than 0.9. However, many products currently on the market fail
to meet these requirements and so more advanced designs are much needed to replace
them in the future.
26 | september 2011 | electronicspecifierdesign
Approaching dimmability...
There are two basic approaches to power factor correction, each of which requires some additional circuitry at the front end of the converter: the simple low cost, passive PFC and the
more complex active PFC. Before exploring
these methods in greater depth, we should mention that in order to gain Energy Star rating, the
LED driver must also be dimmable.
This generally means dimmable from existing
wall dimmers based on the phase cut principle
of operation that originally was designed to
work with purely resistive incandescent lamps.
Although other dimming methods such as linear
0~10V dimming or DALI would presumably
www.electronicspecifier.com
Another Energy Star requirement worth mentioning is that the LED operating frequency has
to be greater than 150Hz to eliminate the possibility of visible flicker. This means that the output
current supplying the LEDs may not include any
significant amount of ripple at twice the line frequency of 50 or 60Hz.
also qualify, they are likely to be limited to high
end industrial type LED drivers.
Phase cut dimmers are by far the most widely
used and it is clear that there would be a significant advantage for LED lamps to be able to be
dimmed effectively by them. Since many low
cost triac based dimmers exist on the market, it
is not practically possible for LED drivers to
guarantee to compatibility with all types, especially since many dimmers are of very basic design and limited performance.
For this reason the Energy Star programme requires only that the LED driver manufacturer
specifies in a web page which dimmers are
compatible with its product.
www.electronicspecifier.com
LED widens its spread
The adoption of LED lighting in off-line applications such as office lighting, public buildings and
street lighting is increasing and is predicted to
continue to do so for the next few years. In
these applications high power LEDs replace linear or high power CFL fluorescent lamps, HID
(metal halide and high pressure sodium) lamps
as well as incandescent lamps.
These applications require an LED driver that
will typically range from 25 to 150W. In many
cases the LED load comprises an array of high
brightness white LEDs often packaged in multiple die form. The dc current required to drive
these loads is often at least 1A. AC current
driven LED systems also exist but dc systems are
generally considered to provide more optimal
driving conditions for LEDs.
In LED light fixtures galvanic isolation is required to prevent electric shock risk where LEDs
are accessible: which is in most cases unless a
mechanical system of isolation is employed. This
is because unlike, for example, fluorescent light
fixtures, these do not need to be isolated for
safety: the LED die need to be connected to a
metal heat sink.
For good thermal conductivity it is necessary
for the thermal barrier between the LED die and
the heat sink. This precludes the possibility of
adding insulating material in between that
would be thick enough to satisfy isolation requirements. It is, therefore, the best option to
provide isolation within the LED driver itself and
this dictates the power converter topologies that
are suitable.
☞
electronicspecifierdesign | september 2011 | 27
driving leds
The two possibilities are the flyback converter or a multi-stage converter that includes
a PFC stage, followed by an isolation and step
down stage, and finally a back end current
regulation stage. Of the two, the flyback is
more popular because of its relative simplicity
and low cost. Flyback converters enable a
good solution for many applications. However,
it has these limitations:
1) Limited power factor correction ability.
2) Limited efficiency over wide input voltage
range.
3) Output ripple at twice the line frequency,
<150Hz, cannot be easily eliminated.
3) Additional circuitry is required for dimming.
Multi-stage design
The multi stage design can overcome some of
these problems, although its additional cost limits its adoption to higher end products. High
power factor and low total harmonic distortion
can be achieved over a wide ac input voltage
range allowing the same LED driver to operate
from a 110V, 120V, 220V, 240V or 277V mains
supply. Efficiency can be maintained over this
range rather than peak at a specific line load
point and drop off significantly under different
conditions. It is also much easier to minimise output ripple under 150Hz and the multi-stage system lends itself more effectively to the different
methods of dimming.
Now we can discuss in detail the design of a
wide input voltage range, isolated, dimmable,
regulated dc output multi-stage LED driver
design concept for applications in the
25~150W range.
The multi stage LED driver in this example will
be broken down into three sections:
1) The front end, power factor correction
section.
2) The isolation and step down section.
3) The back end, current regulation section.
The front end section consists of a boost converter configured as a power factor correcting
pre-regulator that delivers a high voltage dc bus
at the output regulated to a fixed voltage ☞
Fig 1: Flyback converter – simple diagram.
28 | september 2011 | electronicspecifierdesign
www.electronicspecifier.com
Make us part of your electronics line-up.
FIND IT – Quickly.
DESIGN IT – For Free.
BUY IT – In a Click.
rswww.com/electronics
driving leds
over variations in line or load. Since the regulating control loop response is slow and takes
many cycles of ac line frequency to react to
line load changes, it draws an essentially sinusoidal line input current.
Critical conduction operation
This circuit typically operates in critical conduction mode which is otherwise known as
transition mode. In this mode, the PWM off period and therefore the switching frequency is
variable such that the new switching cycle begins at the point when all of the energy stored
in the boost inductor has been transferred to
the output.
This resonant mode of operation is widely
used and provides high efficiency due to minimal switching losses. It is the best approach to
use in the power range required.
The middle stage converts the high voltage
dc bus voltage, typically around 475V, to a
low voltage output suitable for driving LED
loads. For safety reasons LED loads are normally driven by low voltage and thus drive
current is often at least 1A.
The configuration of the isolation and step
down stage recommended here is a resonant
half bridge consisting of a pair of switching
MOSFETs driven in anti-phase with each other.
The midpoint of these switches, supplies one
end of the primary winding of a high frequency step down transformer and the other
end is connected to a capacitive divider network from the dc bus to the 0V return.
In this way, the transformer primary sees a
square wave voltage of equal positive and
negative amplitude. The secondary winding
will be centre tapped so that a two diode rectifier can be used to convert the output back to
dc. Where the output current is sufficiently
high the rectifying diodes can be replaced
with MOSFETs operating as a synchronous
rectification system.
Typical 3A application
In a typical application running at 3A, the
surface temperature of synchronous MOSFETs
was measured at 30°C lower than Schottky
diodes in the same package. It can be seen
that as the current requirement increases, the
thermal benefits of synchronous rectification
become very significant. Finally a smoothing
capacitor is required to produce an isolated
dc voltage with low ripple. This can be in the
order of tens of micro-Farads and therefore ceramic capacitors can be used.
☞
Fig 2: Multistage converter – simple diagram.
30 | september 2011 | electronicspecifierdesign
www.electronicspecifier.com
driving leds
Resonant mode efficiency
For the half bridge stage to operate efficiently,
it needs to be designed to operate in resonant
mode where the MOSFETs switch at 0V. This is
accomplished by ensuring that there is a short
delay between the time when one MOSFET
switches off and its counterpart switches on, and
that during this time the voltage at the midpoint
commutates from one rail to the other.
This takes place due to the release of energy
stored in the inductor conducting through the integral body diodes of the MOSFETs. It is necessary for the primary of the transformer to possess sufficient leakage inductance in order for
sufficient energy to be stored to allow commutation to take place.
Complicated transformer
This makes the transformer design rather more
complicated and one easy way to get around
this difficulty is to use a standard high frequency
transformer design without additional leakage
inductance added into its design and to simply
add another inductor in parallel with the primary solely for commutation.
This extra inductor can also be used to aid
dimming operation from triac based dimmers,
therefore adding justification for the extra cost
and space. This will be further discussed later.
Such an inductor can be built around a gapped or open core to facilitate energy storage.
The back end stage of the LED driver consists
of a current regulating circuit with short circuit
protection. This can be realised with a linear
regulating circuit, however such an approach is
inherently inefficient and therefore only suitable
for low output currents, which will not generally
apply in a multi-stage system. The alternative is
a simple buck regulator circuit with a current
feedback to limit the output current from ever
exceeding the intended LED drive current.
This compensates for variations in total LED
forward voltage over temperature and device
32 | september 2011 | electronicspecifierdesign
tolerance and also limits the current in the event
of a short circuit or some other fault condition
thereby protecting the driver against damage.
A multi-channel approach is also possible
where several output stages are connected to a
single isolated dc voltage supplied by the previous stage. This is advantageous because with
such an arrangement, a short circuit at the output of one of the channels would not prevent
the other channels from working normally. Furthermore it allows several channels of regulated
current to supply different LED arrays and
avoids the need for connecting LED arrays in
parallel. It is well known that connecting LEDs in
parallel is problematical unless the LEDs are of
similar forward voltage drop operating at similar temperature. Thus the advantage of a driver
with multiple independent outputs is apparent.
Triac based dimmer drawbacks
Most dimmers commonly available operate by
leading edge phase cutting, using a very simple
circuit based around a triac. These dimmers
were originally designed to work with incandeswww.electronicspecifier.com
Fig 3: Front end and Half Bridge with
Dimming Charge Pump.
cent light bulbs which are purely resistive loads.
The triac device is a semiconductor switch
that conducts current in either direction between its two main terminals only after it has
been fired by a pulse applied to the third gate
terminal. This pulse can be of either polarity
and is therefore simple to create with a basic
RC timing circuit.
Triac firing
Principle of operation consists of firing the triac
at a point in the ac line cycle so that it conducts
until the end of the cycle at which point the line
voltage drops to zero and consequently so does
the current flowing through the triac. This causes
it to switch off again.
Triac devices have a minimum rated holding
current below which they will switch off. Adjusting a potentiometer in the circuit controls the firing point of the triac in the dimmer circuit and
changes the overall average ac current passed
through enabling dimming.
LED converters and other power supplies or
electronic ballasts, however, do not represent a
www.electronicspecifier.com
purely resistive load to the dimmer even when
they include a power factor correcting front
end. Therefore, the triac in the dimmer tends to
fire erratically and miss cycles when the dimming level is lowered. The factors that influence
this behaviour are fairly complicated and it is
not necessary to go into a deep analysis since a
simple solution has been found which can overcome the problem to a large extent in the multistage system.
DC blocking feedthrough
Instead of returning the commutating inductor
from the primary side of the step down transformer to the midpoint of the capacitive divider,
the current can be fed through a dc blocking capacitor back to the line input. This provides a
small amount of additional current which will
keep the triac from switching off before the end
of the ac line cycle, allowing it to operate as required over the range for dimming. This solution
uses current that would otherwise be wasted to
facilitate dimming using triac based dimmers.
Dimming in this way works because as the
dimmer level is reduced, the output bus voltage
from the front end stage also falls. This results in
the secondary voltage also dropping and since
LED loads have a fixed total voltage drop, a
small variation in voltage causes a large variation in current and therefore light output. In this
way linear dimming for LEDs is realised and this
circumvents the need for more complicated
PWM dimming circuitry. It also avoids possible
patent infringement.
Although dimmer compatibility necessitates
some loss in efficiency, the multi-stage configuration remains a good option for LED driver designs where higher performance is required. ❚❘
Peter B Green is LED Group Manager at International Rectifier.
More from International Rectifier
Return to contents page.
electronicspecifierdesign | september 2011 | 33
showtime
Electronics
Production
On The Up
34 | september 2011 | electronicspecifierdesign
There’s a discernible upwards trend
in electronics production, as
highlighted by the buzz gathering
momentum for productronica
2011 with its new features.
T
he 19th productronica is on the horizon,
opening its doors on November 15 in
Munich. Electronics production is its
theme, as always, for the four day event, with
the links to electronics design perhaps stronger
than ever. And with over 1200 companies displaying their wares in seven show halls, the
event is 15% bigger than 2009.
New developments at the event and certainly helping to maintain the show’s regularly
reported growth, include the PCB Community
Area for the semiconductor industry in Hall
B1, a hall full to bursting.
Positive developments during the preparations for the trade fair, says the organiser, also
indicate that the electronics industry has re-attained the level held before the economic crisis. Registrations are up overall.
“The great demand by companies for productronica 2011,” says Christian Rocke, productronica’s Project Manager, “reflects the
positive market development of the electronics
industry. The pleasing registration figures for
productronica 2011 with much higher exhibitor
numbers and the returning interest of international suppliers exemplify the growth expectations of the market. This is also confirmed by
the forecast in the current business climate survey of VDMA, the German Engineering Federation, in which industry representatives anticipated a 20% increase in turnover in 2011.”
Topical at the show is the organic and
printed electronics segment, based on the
combination of new materials and cost-effective, large area production processes that
www.electronicspecifier.com
open up new fields of application. Thin, lightweight, flexible and environmentally friendly is
the promise of organic electronics which also
enables a wide range of electrical components
to be produced and directly integrated into
low cost, reel to reel processes.
Intelligent packaging, low cost RFID
transponders, rolling displays and OLED lighting, flexible solar cells, disposable diagnostic
devices and games, as well as printed batteries are just a few of the promising fields of application for organic and printed electronics.
This emerging technology is on its path to mass
production and opens up new opportunities for
the electronics manufacturing industry. You can
get up to speed on this at productronica.
PCB community
The new PCB Community Area has attracted
enormous interest, with Hall 1 full and Hall 2
providing an overflow.
This segment is represented
in the Halls by a large number of industry leaders such
as Atotech, Gebr Schmid,
Kuttler Automation Systems,
MacDermid, Posalux and
Schmoll Maschinen.
The concept to
strengthen the European market for PCB
and circuit carrier production in the trade fair
programme is deemed
a success. And the
higher demand in this
arena is also reflected
in other halls at the
show such as placement
technology in Halls A2
and A3 and soldering
technology in A4.
Great interest is also
www.electronicspecifier.com
evident in battery and energy storage device production in Hall B2: also a new segment at productronica 2011. With the related special attraction
Battery Production & Power Electronics jointly organised with the VDMA Productronics Association, productronica is providing a positive focus to
match the trend towards electro-mobility. The key
to electromobility lies in lower costs in production
engineering, the main topic of productronica.
Productronica’s first PCB Community Area is out
to satisfy the huge demand by the PCB industry
for a strong innovative platform to permanently
support the positive developments on the European printed circuit board market. The whole of
Hall B1 is devoted to the PCB segment, providing
exhibitors and visitors with numerous opportunities to obtain information and hold in-depth discussions on circuit board technologies and manufacturing, as well as circuit carrier manufacturing.
The varied programme within the PCB Community Area includes a Networking Area, providing
the opportunity for
intensive and direct
information exchange. A large
number of talks and
discussions con- ☞
electronicspecifierdesign | september 2011 | 35
showtime
cerning current industry topics will take place
during the Speakers Corner Forum and the European Institute of Printed Circuits, EIPC, is also
staging half-day workshops on a variety of related topics. There’s also a special show as part
of the PCB Community Area presenting subareas of circuit carrier and PCB manufacturing.
As reported by the ZVEI PCB and Electronic
Systems Association, the PCB industry is continuing to stabilise again after the severe crisis in
2009. In the past few months, the industry has
experienced substantial growth in sales and subsequently increased its workforce, exhibiting substantial potential on the European PCB market.
This development is also confirmed by Atotech
MD Harald Ahnert who is also a member of the
productronica Technical Advisory Board. “The increase in sales by PCB companies,” he says, “is
a reflection of the high
demand for technological competence from Europe. For companies
such as Atotech, this high
level of competence is attained through close cooperation between research and production.
We are delighted that
productronica has complied with the request by
36 | september 2011 | electronicspecifierdesign
the industry and has developed an attractive concept which appeals to both exhibitors and visitors
from the industry and will bring them together.”
“The idea of the PCB Community Area,” explains Christian Rocke, “is that printed circuit
boards have always been one of the most important and most central topics at productronica.
Established and recognised throughout Europe
as the leading trade fair, productronica is an
ideal platform to strengthen companies and,
thus, the market. With the separately created
marketplace, we will present the European PCB
industry in its entirety.“
Battery & power
The battery production and power electronics
section within productronica is playing an important role in electromobility where many produc-
www.electronicspecifier.com
tion engineering challenges remain unresolved.
“The acceptance and thus the economic success of electromobility worldwide depend to a
very large extent on the further development
of battery technology,” emphasises Rainer
Kurtz, Chairman of the Kurtz Group Board
and also Chairman of the VDMA. “It is quite
rightly a key topic in the interim report of the
German Federal Government’s National Platform on Electromobility, which was recently
published. High performance energy storage
devices must not only become more efficient,
but cheaper above all. This objective will only
be achieved with highly developed production
engineering. A massive cost reduction will only
be attained through highly integrated production and interaction among all players along
the process chain.”
This special show within productronica 2011 will link battery production
for electromobility to the topic of
power electronics which is indispensable for battery packaging and all
other high energy components ranging from inverters through to the drive
train in automobiles.
“Only mass production will lead to
learning curve effects and thus lower
costs,” maintains Dr Eric Maiser,
VDMA Productronics MD. “Computer
chips, flat panel displays and photovoltaics clearly show that technological
developments are possible alongside
with enormous cost degression. The German engineering industry has a great deal of experience on these markets and production engineering has a number of similarities which will take
us forward more quickly. As the platform for the
electronics production industry, productronica is
an ideal setting for presenting and linking these
capabilities. During the special show we will
demonstrate what is required in manufacturing
high performance energy storage devices ☞
www.electronicspecifier.com
Name
Dr. Christian Altenbach,
Certi ed LabVIEW
Associate Developer
Job Title
Research
Ophthalmologist
Area of Expertise
Biophysics
LabVIEW Helped Me
Analyse and visualise
data interactively using
custom algorithms
Latest Project
Mapping molecular
structure changes
during activation of
the light receptor,
rhodopsin
NI LabVIEW
LabVIEW makes me better because I can
the way I think
>>
Find
Find out
out how
how LabVIEW
LabVIEW
can
can make
make you
you better
bet ter at
at
ni.com/labview/better
ni.com/labview/better
01635
01635 517300
51730 0
uk.ni.com
uk.ni.com
info.uk@ni.com
THE
CONVERSATION:
Follow
us on
on
JOIN
CONVERSATION:
Follow
JOJOIN
IN THE
CO
NVERSATION: F
ollow us
Search National Instru
Search
Search National
National Instruments
Instruments or LabVIEW
LabVIEW
Austria 43 662 457990 0 Q Belgium 32 (0) 2 757 0020 Q Czech Republic, Slovakia 420 224 235 774
Denmark 45 45 76 26 00 Q Finland 358 (0) 9 725 72511 Q France 33 (0) 8 20 20 0414
Germany 49 89 7413130 Q Hungary 36 23 448 900 Q Ireland 353 (0) 1867 4374 Q Israel 972 3 6393737
Italy 39 02 41309277 Q Netherlands 31 (0) 348 433 466 Q Norway 47 (0) 66 90 76 60
Poland 48 22 328 90 10 Q Portugal 351 210 311 210 Q Russia 7 495 783 6851
Slovenia, Croatia, Bosnia and Herzegovina, Serbia, Montenegro, Macedonia 386 3 425 42 00
Spain 34 (91) 640 0085 Q Sweden 46 (0) 8 587 895 00 Q Switzerland 41 56 2005151
UK 44 (0) 1635 517300
©2011
National
NI,
© 2 0 11 N
at ional IInstruments.
ns t r umen t s . All
A ll rrights
igh t s rreserved.
eser ved. LabVIEW,
L abV IE W, National
Nat ional Instruments,
Ins t r umen t s , N
I, and
and nni.com
i . co m
are
National
are trademarks
t rademar ks ooff N
at ional IInstruments.
ns t r umen t s . Other
O t her pproduct
roduc t aand
nd company
company names
names llisted
is ted are
are ttrademarks
r ad e m ar k s
oorr trade
t rade names
names ooff their
t heir rrespective
espec t i ve companies.
companies . 22785
785
electronicspecifierdesign | september 2011 | 37
showtime
with respect to production engineering aspects
and how the industry is now already getting
ready to do this.”
High energy storage devices and power
electronics are important not only for mobile
but also for stationary applications too. “Energy generation from renewable sources,”
says Kurtz, “not only calls for efficient transmission grids but also capacities for intermediate storage. Battery storage devices can make
a major contribution, in particular, to decentralise energy supply.”
Maiser concludes, “With this, we are also
addressing a [good] interrelationship to photovoltaics production here. The main requirements for all these objectives are cost-efficient
and quality assured production technologies,
the main theme of productronica.”
Trending vibes
Among underlying trends, in semiconductors
and inspection equipment, for example, electronics manufacturing seems on track for better things once again. According to the IHS
global forecast, production of semiconductors
on 12in wafers will nearly double by 2015.
"Initially, 12in wafers were employed only for
the most advanced products," says IHS’s Research Director Len Jelinek. "However, that’s
been changing over the course of the past two
years, with foundries and IDMs having determined that 12in wafers represent the most costeffective manufacturing method for mature
products. As a result, IHS forecasts a new period of rapid growth for 12in wafers."
According to Frost & Sullivan, the SMT inspection equipment market is also poised for growth,
with smartphones to tablets and convergence in
the connected home the buzz of electronics.
Change seems to be the new mantra, says the
research organisation, and OEMs are focusing
on creating value through technology. Consequently, the proliferation of miniaturised components is on the rise, along with
denser board packages.
Denser and complicated assemblies, which include BGAs, flip
chips and chip scale packaging
need automated inspection to ensure quality and hence we see
Frost & Sullivan’s vision for a new
drive in surface mount technology
inspection equipment.
Whether fact or fantasy, productronica will certainly give a good
marker to trends in electronics emanating from electronics produc❚❘
tion demand.
Return to contents page.
38 | september 2011 | electronicspecifierdesign
www.electronicspecifier.com
grass roots emi
Meeting EMI For AC/DC Systems
E
Vicor Westcor’s David Fletcher takes us back to filter basics.
MI is often seen as one of the black arts
in power conversion design and system
integration, particularly for ac/dc systems. Here we look at the causes of EMI and
its impact on a system, before going on to
make some practical suggestions on how to
minimise its effect.
Causes of noise
We have to start with what
engineers affectionately call
noise. This is generated
whenever rapid voltage,
dv/dt, and/or current, di/dt,
transitions take place.
AC/DC power converters
typically utilise a number of power switching
topologies including
fully resonant, quasi-resonant and PWM, the most typical control in the front end section
of the converter.
PWM controlled converters use
a rectangular control signal with a
continuously varying pulse width
in response to the operating conditions of the converter. The result is
typically a white noise energy distribution spectrum. And if this is not filtered and shielded, it interferes with
consumer electronic equipment using the same
ac mains.
Noise currents exiting the power converter
through the ac power lines and/or via output
power cables are known as conducted emissions. The noise manifests itself in two forms:
www.electronicspecifier.com
differential and common mode.
The definition of differential mode is noise
that is only on the power lines and is not present on the earth ground lead, and it can be
measured with respect to the power lines. The
definition of common mode is noise that can
only be measured from earth ground to
one of the power lines.
AC/DC power converters employ EMI filters within the
power converter enclosure.
These filters incorporate noise
suppression topologies containing inductive and capacitive components called X
and Y elements.
X components are
placed across the
power lines and filter the differential mode noise. The
Y components are placed between
the power lines and earth ground
and filter common mode noise.
Minimising EMI filtering
AC/DC power converters are designed to meet various regulatory and
safety standards, including various
EMI standards. The individual details
and the standards met are generally
covered thoroughly in available
product literature.
However, during a system integration of various components, including an ac/dc power
converter, a system design engineer may find it
necessary to add additional EMI filtering to enable the integrated system to comply with ☞
electronicspecifierdesign | september 2011 | 39
grass roots emi
all relevant agency standards.
If this approach becomes necessary, it is recommended to minimise the additional filter components required by empirically gathering test
information. It is recommended that a few areas
be given special attention:
❏ Test the system without any extra filters. This
❏ The filter must provide enough attenuation
at the noise frequencies of concern to provide
enough margin for system to system manufacturing variations.
❏ The amount of extra leakage current or
will give the system engineer a good baseline.
❏ Identify the problem frequencies on the EMI
plots.
❏ Try to relate these frequencies to areas of
the integrated hardware within the system.
❏ Re-check all points of earth grounding and
cable routing. All these should be in accordance with good engineering practice. These
include, but are not limited to, separating signal cables from power cables, using twisted
pair techniques of 3~5 twists/in if possible, on
cable carrying signals and/or high current
and their ground returns.
If using this technique is deemed impractical
then strip lining the cable as close to one another is an alternative to minimise any noise.
Consider shielding cables with fast voltage
and/or current transitions.
In integrated systems that have several
grounding points a careful study of the system
integration block diagram is advisable to detect the presence of grounding loops. If these
are present they must be eliminated.
Adding external filtering considerations
If EMI problems still persist, then adding an
external filter may be necessary. There are
areas to consider when considering an off the
shelf filter:
❏ The filter must be able to handle the full
rated, worst case current of the system.
40 | september 2011 | electronicspecifierdesign
Block diagram showing differential
mode and common mode noise.
www.electronicspecifier.com
touch current caused by the additional Y capacitance inherent in the filter should be calculated. If the leakage current is not stated on
the filter data sheet, the following formula can
be used to calculate it:
2 x π x (Vac max) x (ac frequency max) x
(Y capacitors) = I leakage
Some 20~25% should be added to this
value before you add this to the overall leakage current. If the total leakage current exceeds the allowable regulatory
level, you may
want to consider a
filter that has an
internal series inductive component in the
ground lead.
These filters tend to have no Y capacitors
and may be referred to as medical filters.
Care should be taken to prevent the ring-up
phenomenon which occurs when the forward
current through a series inductor is high
enough to saturate the filter inductor so that
when the resultant magnetic field collapses, a
very high peak may occur, with a voltage amplitude up to twice the original impulse.
These peaks may be caused by lightning
strikes, external power company grid switching or heavy inductive transitions due to motor
loading in a building.
Commercial converters have input protection
schemes and therefore comply with
regulatory standards.
However with the addition of an external filter
through ring-up,
it is possible to
exceed
❏ The filter should have an internal resistor
the protection components and
cause damage to the power converter. External protection components, such as MOVs
may be needed across the external filter to
add robustness to the overall system.
❏ The effect of adding an external filter
ble to the ac input of the system enclosure to
maximise EMI filtering.
All ground connections to the filter should be
as short as possible to minimise series induc❚❘
tance and impedance.
between the power lines to ensure that on disconnection of the ac line input, the internal X
capacitors discharge properly: typically
within 1s, as recommended by VDE 0806
and IEC380.
means additional inductive elements are now
in series with the power converter’s ac input.
www.electronicspecifier.com
❏ The filter should be placed as close as possi-
More from Vicor
Return to contents page.
electronicspecifierdesign | september 2011 | 41
cabinets, racks & enclosures
P
Containing ESD
Paul Hoath from Vero Technologies checks out
coating finishes for moulded enclosures.
lastic enclosures have many benefits.
There are many suppliers serving the
market with a vast number of standard
products available in different styles and
sizes at economical prices, reducing the
time to market for new projects. Products
are readily available from manufacturers and also from
broadline and specialised distributors.
The units are lightweight and available
in complex shapes.
Features are easily incorporated into designs and they are
readily modified to
accept displays,
switches, connectors
and other hardware.
potential problem, the lack of screening may
well be an issue.
Enclosure manufacturers have addressed this
issue, typically by adding conductive coatings
to the inside of the enclosures, although there
are some products on the market that are
moulded from conductive plastic.
Whichever approach
is used – and the
coating method is by
far the most popular
– the design of the
mouldings can have
a significant impact
on the screening effectiveness of the
conductive coating.
Most plastic enclosures are constructed
from a top and base
moulding: in some
designs, there may
also be battery compartments and removable end panels.
EMC challenges
However, moulded
enclosures have one
specific shortcoming.
By virtue of the intrinContinuous surface
sic properties of the
One key requireenclosure material,
ment for effective
plastic, unlike metal,
offers no inherent at- Vero has a wide range of standard plastic enclosures. screening is that the
internal surface of
tenuation to the passage of electric or magnetic fields. In many ap- the enclosure should be as continuous as posplications, this deficiency is of no consequence sible, to ensure electrical conductivity between
but if radiation emitted by the housed electron- all its parts. In particular, long slots should be
avoided. To prevent slots between the mating
ics or their susceptibility to external fields is a
42 | september 2011 | electronicspecifierdesign
www.electronicspecifier.com
halves of a typical enclosure, tongue and
grooving provides an effective complex path,
improving attenuation performance.
Obviously, the higher the frequency, the
lower the wavelength, so even very small gaps
can have a detrimental effect on EMC performance.
If there are removable end panels, they need
to be secured via an interference fit slot or, if
they are secured using fixings, a conductive
gasket should be fitted to the mating surface.
The rear face of plastic panels needs conductive coating, or if aluminium panels are used,
the front surface will normally be anodised
and the rear left with a natural finish or
iridised, a RoHS compliant conductive finish.
The best methods of providing suitable continuity in enclosures with dedicated battery
compartments will depend on the design.
If the battery box is constructed with solid partitions between it and the main internal space of
the enclosure, the only precautions that need
be taken are to ensure that the hole for wires
entry into the enclosure is as small as possible.
If the battery is just clipped into mouldings in
the enclosure without a partition, then the lid
of the battery box will form part of the overall
screening and will be a potential screening
weak spot as there will typically only be a
flat surface interface with the body of the
enclosure itself.
Sprayed on attenuation
Conductive coating suppliers have developed several different main materials for
spraying the inside of enclosures to achieve
different levels of attenuation versus cost.
Vero Technologies, working closely with its
supplier partners, R F Solutions and Polymer
Coatings, provides three alternative coatings
that cover more than 95% of possible applications and for highly specialised uses, other
coatings are also available.
www.electronicspecifier.com
Nickel coating
Meeting general commercial level requirements, a nickel based colloid delivers acceptable attenuation at a competitive cost with a
50µm thick film.
Figure 1: Nickel coating performance graph.
The material is COSHH, RoHS and REACH
compliant, providing attenuation of 60~65dB at
50µm when tested to ASTM ES7-83 which is a
standard test method for measuring the electromagnetic shielding effectiveness of planar materials. After temperature ageing of seven days
at 29.4°C, 95% RH, no property degradation
was found following environmental testing.
Copper coating
For more severe requirements, a copper colloid, formulated with silver coated copper particles and conductive resins, provides a conductive layer. It provides an effective shield
against RFI and EMI and can act as a ground
plane to protect against ESD.
As can be seen in Figure 2, page 44, high
frequency performance is better than that of
nickel. Inevitably, given the relative costs of
the base materials, the higher attenuation
costs more.
While not shown on this graph, the material
has been tested up to 10GHz to Mil Std 285,
typically providing 78dB at 10GHz. The material is COSHH, RoHS and REACH compliant.
After temperature ageing for days at
☞
electronicspecifierdesign | september 2011 | 43
cabinets, racks & enclosures
Figure 2: Copper coating performance graph.
85°C, 85% RH and ten temperature cycles of
75°C for 1h, then 1h ambient and –30°C for
1h, followed by 56 days of high humidity testing at 35°C, 95% RH, no degradation in attenuation properties was found.
Striking the optimum balance between cost
and performance, the copper colloid is Vero’s
default coating material, although higher or
lower performance coatings can be specified
as required.
Silver Coating
Highest attenuation achieved by the Vero
coating trio is its silver colloid, formed of silver
flakes and conductive resins. It provides an effective shield against RFI and EMI and can act
as a ground plane to protect against ESD.
Figure 3: Silver coating performance graph.
This material is COSHH, RoHS and REACH
compliant. After temperature ageing for days
at 85°C, 85% RH and ten temperature cycles
44 | september 2011 | electronicspecifierdesign
of 75°C for 1h, ambient for 1h and –30°C
for 1hr, then 56 days of high humidity testing
at 35°C, 95% RH, the attenuation properties
were found not to degrade.
Many applications require a display in the enclosure, normally viewed through a transparent
window. And two main alternatives are available to preserve the integrity of the internal conductive coating when a window is required.
A wire mesh will provide continuity at the expense of clarity. The better option, originally
developed for use on military helicopters, is a
clear conductive coating that provides the required electrical conductivity without obscuring the display. Vero can apply the clear coating to any window material as required.
ESD Protection
Electro-static discharge can be minimised by
applying graphite or carbon-sed high conductivity coatings. ESD events can occur without a
visible or audible spark at low voltages
around 10V and this is sufficient to damage
sensitive electronic components. They can
cause outright failure or reduced longterm reliability and performance reduction.
External coatings
Several different types of coating can be applied for specific applications. Whilst they are
obviously mutually exclusive, for medical use
anti-bacteriological coatings inhibit the growth
of e-coli, MRSA and other bacteria.
Fire retardant, low smoke coatings can be
applied in aerospace applications. Protective
coatings that resist harmful substances found in
petrochemical, pharmaceutical and similar industries are available, and low friction coatings reduce surface wear. High visibility florescent, luminescent and iridescent coatings have
benefits in safety critical applications.
So, moulded enclosures are a popular
choice of housing. They are available in sizes
www.electronicspecifier.com
from key fobs, through handhelds to desktop
instrument housings, providing style and functionality at low cost. And with the addition of
high performance coatings to the internal and
external surfaces of these units, they can also
provide enhanced protection against damage
to the housed electronics from a wide range of
❚❘
external threats.
(Graphs courtesy of RF Coatings Ltd.)
More from Vero
Return to contents page.
Boxing Clever…
IBX is a desktop or mobile enclosure with carry
handle that accepts single and double Eurocards
Vero Technologies manufactures the widely used in standard or extended depths. These modules
plastic enclosures originally developed by Vero
can also be mounted in vertical or horizontal
Electronics and acquired from the receivers of its configurations. IDAS is a desktop modem case
successor company, APW. Brands such as
and instrument case for single and double EuroVeronex, IDAS, Apollo, General Purpose Box,
cards in various heights in 35mm increments. It
Patina and many others continue in full scale pro- can be fitted with a carry handle and a range of
duction, enabling companies who had previously accessories is also available.
specified the products as the housings for their
Patina accepts single and double Eurocards up
equipment to continue to purchase and design-in to 100mm high for vertical or horizontal mountfor new projects. All products are available exing. It can be desk or wall mounted and features
stock from VeroDirect, the company’s 24/7 web a traditional or clip fixing assembly. Apollo deskshop, www.verodirect.com.
top or wall mounted enclosures accept cards up
The Veronex family comprises handheld cases, to 130x190mm. The General Purpose case is
instrument cases, IP64 rated and EMC RFI
handheld for 55x35~85x155mm PCBs.
shielded enclosures, suitable for a range of card
Flip Top cases are designed around a 100x
sizes with both vertical and horizontal mounting 160mm standard depth single Eurocard and the
and a wide range of options and accessories.
Desk Top case accepts PCBs up to 90x135mm. ❚❘
www.electronicspecifier.com
electronicspecifierdesign | september 2011 | 45
cabinets, racks & enclosures
The Third Certainty...
Schroff’s Marcus Edwards offers his thought on energy costs
and how they fit into Franklyn & Bullock’s definition.
I
Schroff’s cold aisle containment solution.
n this world, nothing is certain but death
and taxes! This well known quotation is
usually attributed to Benjamin Franklin
from 1789, although it should be ascribed
to Christopher Bullock, the playwright, who
first used it in 1716. However, some 300
years on and the statement has lost none of
its relevance but it can now be brought up
to date with the addition of the words ‘and
rising energy costs’!
For the data centre owner, electricity costs
are the largest single overhead and, as every
indicator points to the unit cost/kWh increasing remorselessly with time, any steps that can
be taken to control energy usage should be investigated. Input power is used in two ways: to
run the servers, routers, system management
and other electronic equipment, and, since
nothing is 100% efficient, it is needed to re46 | september 2011 | electronicspecifierdesign
move the waste heat generated.
The accepted measure of a data centre’s efficiency is power usage effectiveness, PUE, defined as the ratio of the total power consumed
by a data centre to the power consumed by
the IT equipment itself.
In a ‘traditional’ CRAC cooled data centre, the
PUE will typically range between 3, in an inefficient facility where for every kW used to power
the equipment, 2KW is used for ancillary functions and, in the most efficient operations, 1.2,
where 83% of the energy is used for the primary purpose of powering the IT equipment.
Many alternatives
Many alternative approaches can be taken to
improve the PUE of a data centre. Incidentally,
careful measurement at regular intervals during
the day over an extended period are required
www.electronicspecifier.com
to arrive at an accurate measure of PUE, which
varies according to the time of day, the server
loadings, the season of the year and several
other fairly obvious variables.
In addition to the main problem of bulk heat
removal, effective thermal management is critical to the reliable operation of IT equipment.
Some 55% of all equipment functional and
catastrophic failures are attributable to thermal overload.
Both issues are addressed in the first step,
basic air management, which must be implemented before more sophisticated solutions
can be considered.
Hypothesised measures
To quantify the impact of the various measures that can be taken, let’s consider a relatively small hypothetical data centre containing 40 cabinets, each using 5kW for the
housed IT equipment. The cabinets are
arranged in four aisles of ten, each in a hot
aisle/cold aisle configuration.
The room itself is of traditional construction
with overall cooling provided by a CRAC system, with the cold air distributed from one end
under the raised floor on which the cabinets
are mounted. The energy consumption in such
a non-optimised scenario is estimated to cost
£484k pa with a poor PUE of 2.75.
Three improvements can immediately be implemented at insignificant capital costs.
Blanking panels inserted in unused positions
within the rack will prevent recirculation and
reduce the temperature gradient between the
bottom and top of the rack, helping to manage hot spots.
Floor brushes on the cable entries in the tiles
under the cabinets will stop cold air leakage
through the base, forcing the airflow up
through the cold aisle in front of the racks.
Of greater significance is the positioning of
the ventilated floor tiles to ensure that the unwww.electronicspecifier.com
derfloor cold airflow is as uniform as possible
throughout the room.
Cabinet over-heating
High linear airflows close to the CRAC outlets can result in overheating of the nearest
cabinets, as the air is moving too fast to go up
through the ventilated floor tiles. Conversely,
those cabinets furthest away from the CRAC
rely on sufficient airflow to ensure that they
are not starved of incoming cooling air.
A delicate balancing act, which inevitably involves compromises, can achieve the optimal solution. Louvered ventilated tiles enable the airflow
to be adjusted as required. In our notional data
centre, the PUE will improve slightly to 2.65, saving some £16k, or 3.7% in energy costs.
Of greater importance is the improvement in
equipment reliability as the result of the more
uniform airflow through the enclosures. But it
must also be stressed that without good practice in the fundamentals of air management,
additional, more complex changes will be notably less effective.
Smooth ventilation
Ventilated front and rear doors smooth front
to rear airflow within the cabinet. These will
typically contain the airflow to allow up to
7kW heat generation per cabinet.
Other cabinet level options include additional exhaust fans to improve the volume of
air moving through the cabinet, although fans
are a known failure point that could compromise equipment life and operational reliability.
Other options in an open data centre include
rear door mounted, air/water heat exchangers that efficiently remove additional heat from
the servers and which, if suitably sized and
controlled, will also reduce the temperature of
the exhaust air returned into the data centre.
Rear heat exchangers will allow a load of up
to 15kW per cabinet.
☞
electronicspecifierdesign | september 2011 | 47
cabinets, racks & enclosures
Chimney ducts, in which the hot exhaust air
is vented directly into the CRAC ducting has
been tried, but it is an expensive option in
terms of infrastructure requirements and requires complex design and planning. The overall impact on energy consumption is limited.
Defacto choice
Hot or cold aisle containment is the de facto
choice for improving thermal efficiency of the
data centre. This typically enables a 20~30% reduction in energy consumption in return for limited capital investment, giving ROI times typically
measured in months rather than years.
Roofing the cold aisle and fitting
doors to each end gives
total separation between the hot
and cold air
in the
room.
This allows
over-pressuring
of the input air, minimising vertical temperature
gradients, thus enabling higher
packaging densities within the cabinets.
Upgrade example
A good example of the effectiveness of cold
aisle containment recently involved Schroff in
an upgrade project. The data centre contained
75 server cabinets arranged in three rows: the
CRAC system comprised five 32kW units, all
running at 100% capacity and the vertical
delta T within the cabinets was 10°C.
When cold aisle containment was fitted and
airflow optimised inside the cabinets the delta
T reduced to 1.5°C, the room temperature declined by 6°C and the delta T between the
CRAC output and the cold aisle was almost
non-existent.
The outcome was that the cold aisle tempera48 | september 2011 | electronicspecifierdesign
ture could be raised from 17 to 23°C and the
output temperatures increased from 22 to
30°C. The CRAC fan speeds were reduced
and one of the units completely shut down,
saving some 30% on energy bills.
Aisle containment is the most cost-effective
means of improving a data centre with a false
floor and central CRAC cooling. Cold aisle
containment can be reinforced by the use of
an in-row chiller, either as a standalone unit or
as reinforcement to a CRAC system.
Data centre cooling schematic, left.
New builds
However, with a
blank piece of
paper in a new build
facility, overhead
cable distribution
removes the
need for a
false floor
and intelligent
closed loop liquid heat
exchangers fitted to sealed
server cabinets create controllable
micro-climates, enabling up to 40kW to be
removed from a single cabinet. A high capacity CRAC system is not needed to cool the
room. Instead, the normal building air conditioning system will be adequate.
Energy usage reduction is typically up to
35%, and as an additional benefit, infrastructure investment is almost non-existent. Hot aisle
containment and in-row cooling is an alternative approach for new build facilities and compared with the non-optimised data centre, savings of 45% can be achieved as the PUE
improves from 2.75 to 1.50.
A PUE of 1.5 is arguably the best that can be
achieved by changes to the data centre configuration and cabinet level changes. To achieve a
PUE of 1.2 in a retrofit, the installation of evapowww.electronicspecifier.com
ration chillers or direct air cooling or other very
capital intensive changes will be needed.
In new build facilities, closed loop integrated liquid heat exchangers enable data centres to be
created in simple buildings that do not need comPayback inevitability
plex and expensive
The inevitable result is
CRAC systems and
that the payback time
raised floors. This reon the capital invested
duces investment and
will be considerably exenables the facility to be
tended. In green field
brought on-line as
developments in Northquickly as possible.
ern Europe, it is most
Intelligent and scalalikely that only a relable thermal managetively small CRAC unit
ment is the most flexible
will be installed to proand cost-effective mechSchroff’s hot aisle/cold aisle solution.
vide additional cooling
anism yet developed to
on the few hot days in
reduce costs, reduce
the year when the external air/water chiller is un- time to market and cope with ever-increasing
able to cope with demand.
packaging densities and higher power usage
So, in conclusion, given the inevitable on-going within multi-core processors in space efficient
escalation in energy costs, improving PUE is criti- blade servers. ❚❘
cal. ROCE is generally excellent, often measured More from Schroff
Return to contents page.
in months rather than years.
IT Cabinets
VariStar server and networking cabinets from
Schroff provide future-proof solutions for the IT
sector, especially in data centres where traffic increases and applications are becoming increasingly more complex.
Scenarios include quickly reached server capacities and inevitable system expansion. Demands on energy, cooling and safety increase
accordingly. But that is no problem, says Schroff,
provided the infrastructure is flexible enough to
keep pace with changing demand.
VariStar comes in a raft different versions and
with a comprehensive range of accessories, with
full compatibility of components from system to
system. The solution makes for easy and straightforward expansion.
The cabinet system features power, current
and voltage monitoring of individual sockets,
www.electronicspecifier.com
enabling data analyses directly or via the web.
Remotely monitored and controlled socket
strips and programmable On/Off switchers
save operating costs and time. The system also
features advanced authorisation options with
256bit AES encryption and in operation, costs
and time can be saved with remote power cycling and monitoring.
Power consumption measurement and management built into VariStar make for greater system
efficiency, tracking down power heavy applications and system devices with large thermal
loads. Power control strips measure power at the
socket, providing specific data for energy optimisation. Individual outlets can be switched on or
off remotely enabling fewer key applications to
run in critical situations. And with sequential
switch-on of each outlet, high start-up currents
❚❘
are no longer a problem.
electronicspecifierdesign | september 2011 | 49
enclosures
Standard Boxes Clever For Simplicity
S
Hammond’s Russell Irvine makes the case for modified
standard over custom enclosures.
tandard enclosures for electronic and
electrical applications come in a huge
variety of sizes, styles and materials,
from many specialist manufacturers. And the
initial choice facing the designer is whether
to commission a custom enclosure for the
particular project or whether to choose a
suitably sized standard one from a reputable supplier.
The main advantage of standard enclosures is
that there are no recurrent engineering and
tooling charges. Moreover, most manufacturers
keep products in stock so the time to market is
as low as it can be and unit costs are attractive
and the design has been field proven in many
different applications.
Obviously, there may have to be compromise in
the mechanical design of the PCB and other
components to ensure they fit into the enclosure but, providing the enclosure is selected
at an early stage of the design process, this
will not generally be a problem. In most
cases, the PCB will be of a custom size and
shape anyway and it is easier to make the PCB
fit the housing than it is to find a slightly larger
version of a standard enclosure.
Choice of material is an important factor in selecting enclosures. Ignoring metal enclosures,
which are normally built from an aluminium extrusion framework with sheet aluminium or steel
cladding, the two main materials used for formed
enclosures are die-cast aluminium and plastic.
Die-cast aluminium housings are strong and robust. They do not corrode, are electrically conductive, have an intrinsically high level of elec50 | september 2011 | electronicspecifierdesign
tro-magnetic attenuation and are easily machined. With a suitable gasket fitted between
the lid and the base, environmental sealing to
IP67 can be easily achieved, and by adding inserts during the casting process, repeated openings and closures are facilitated.
Such enclosures can be cast with relatively thin
walls, although they
will always be far
heavier than the
equivalent sized
plastic moulded
one. For applications where protection against shock
damage is im-
portant,
where
EMC is likely
to be an issue or where
high temperatures, dust or water
are expected to be present, the die-cast enclosure is the ideal low cost choice.
Thermoplastic moulded enclosures are typically produced from either polycarbonate or
ABS. Both are lightweight, are good electrical
insulators and can be moulded to extremely
fine tolerances to give good detailing on the
finished product. ABS is considered superior
www.electronicspecifier.com
for its hardness, gloss, toughness, and electrical insulation properties but it is more expensive than polycarbonate.
Both materials exhibit good shock resistance
and are shatterproof. Polycarbonate has the advantage that it can be moulded in clear or
translucent colours, enabling, for example, internal displays to be seen through the case itself or
infrared emitters used in remote control applications to be internally mounted and shine through
the polycarbonate material. As with die-cast aluminium enclosures, environmental sealing, typically up to IP65, can be achieved though the
addition of a suitable gasket material between
the apertures and the body of the unit.
Whatever style
and type of standard
enclosure is specified, it inevitably will
have to be adapted
to accept the
switches, LED indicators, dials, keyboards and connectors required by the
application. Typically, the enclosure
will
have to be
machined with
suitable apertures to
accept the various components, and it will probably have to
be silk screened with legends and logos. Diecast enclosures may also have to be painted.
For these needs, the purchaser has three
choices. Standard enclosures can be purchased and subsequently modified in-house:
www.electronicspecifier.com
modifications can be carried out by sub-contractors or the enclosure can be modified by
the manufacturer and supplied as a finished
unit ready for the components and PCB to be
assembled in to it.
Depending on in-house, the manufacturing
capabilities, to modify the housing as part of
the assembly process may be the best way to
proceed, although in most cases, more standard units than are actually required will have
to be purchased to allow for set-up procedures
and wastage. The same holds true if the modification processes are outsourced but in this
case there are the additional costs and logistics complications associated with managing
the sub-contracting process, often to more
than one outsourced operation.
By far the best option is for the original manufacturer to provide the modified enclosure. By
working with the manufacturer, the user will ensure that there are no weak spots because, for
instance, holes have been drilled so closely together that there is not enough land to give sufficient strength when the switches or connectors
are used. If a plastic enclosure is required in a
special corporate colour, it is far better and
cheaper to mould it in that colour rather than
paint it and if the application requires specific
mechanical properties from the enclosure, they
can often be supplied by using a special blend
of plastic or additives.
A superficial glance at a moulded or die-cast
enclosure may give the impression that it is a simple box. But in reality it is the end result of a
great deal of design effort and expertise that has
been expended to create the most useful and
feature rich housing, suitable for a use in a wide
variety of applications and environments. Work
with the enclosure manufacturer from an early
stage in the project design cycle to tap the wealth
of experience behind a standard enclosure! ❚❘
More from Hammond
Return to contents page.
electronicspecifierdesign | september 2011 | 51
powering fpgas
Powering Down FPGA Power
Supply Component Count
A
New µModule regulators from Linear Technology are out to reduce the power
supply component count for FPGA based systems, says Afshin Odabaee.
lthough the versatile and configurable
nature of FPGAs are attractive to system designers, the complex nature of
the design rules governing the inner workings
of these devices and their outer interface protocols, require extensive training, reference design evaluation, design simulation and verification. As a result, FPGA suppliers provide
detailed hardware and firmware support to assist system architects grapple with new challenges in the digital domain.
However, obscure intricacies in the analogue
52 | september 2011 | electronicspecifierdesign
domain, specifically in the realm of delivering
power and regulation voltages with dc/dc regulators for core, I/O, memory, clocks and other
rails, demand new solutions. For instance,
today’s FPGAs and supporting components require multiple voltage rails. To power each rail
efficiently and in the smallest space possible requires a dc/dc regulator circuit that contains, on
average, ten components, including inductor,
MOSFETs, capacitors and dc/dc regulator.
A six rail FPGA may need as many as 60 components just to power it. Aside from a long list of
www.electronicspecifier.com
components required to power the FPGA, there
are hidden costs in component insertion, reliability, PCB complexity and more. Now it’s time for
dc/dc manufacturers to raise the performance
bar of their products!
Managing multiple voltage rails
Older generation FPGAs required two or three
power rails. Now some of the high end multiple
core devices require as many as seven, a mixture of 3.3V legacy power rails and recent
lower voltages ranging from 2.8V down to 1.0V
and less. Moreover, there is a mix of other voltage rails provided for devices other than an
FPGA: memory; network processors; graphic
processors; DACs or ADCs, as well as op amps
and RF ICs.
Ensuring clean start-up in a system with multiple voltage rails, without any of the rails conflicting with one another is the crucial task of a
dc/dc regulator equipped with sequencing and
tracking functions.
Simply stated, each regulator must be able to
track the output voltage of other regulators. The
good news is that for a few years now, FPGAs
have no longer required rail sequencing. However, sequential ramp up or ramp down of several voltages across different sections within a
system are still required to prevent possible
latch-offs that may occur when a voltage rail
comes up too fast or too slow.
In the past, the tracking and sequencing of
power rails fell on a separate power management IC. Today, designers require that both the
sequencing and tracking functions be embedded
into the regulators, especially when they must be
located at different corners of the system.
Lowering voltage ripple noise &
capacitor requirements
In non-portable applications, as the requirements for voltages drop and currents rise, heat
dissipation and operating efficiency become
www.electronicspecifier.com
more important factors in the selection of a
dc/dc regulator. In portable applications, although load current per rail is less, the operating and standby efficiencies remain as important factors in preserving a battery’s energy
and simplifying thermal management of the
portable product.
A switch mode dc/dc regulator enables a
higher performance solution than a linear regulator, especially in either portable or nonportable applications, for high power requirements. For example, a switch mode regulator
providing 1.2V at 5A from a 3.3V input supply
at 90% efficiency compares to a linear regulator’s 36% efficiency. Furthermore, whereas the
switch mode regulator dissipates 0.7W, the linear regulator is dissipating 10.5W.
On the other hand, a switch mode regulator
introduces switching noise and higher output ripple noise – output voltage peak to peak ripple
– because of its inherent switching operation.
Unfortunately, the lower voltage rails of new
FPGAs and tighter eye diagrams of faster I/O
signals are less tolerant of power supply noise.
To alleviate the ripple noise, more input and
output capacitors can be added to the circuit to
dampen peak to peak ripple voltage. However,
damping switching noise is more challenging.
One possible approach is to synchronise the
dc/dc regulator’s operating frequency to an external clock which forces the regulator to operate within a set frequency chosen to have minimum interference with other noise sensitive
parts of the system. This method is especially effective when several switch mode regulators are
all synchronised to a clock frequency that is safe
to the rest of the system.
These methods help with the design of a less
noisy switch mode point of load regulator. However, the problem of noise can be greatly reduced if the dc/dc regulator is designed from
the ground up with the proper architecture,
functions and layout. Such a regulator mini- ☞
electronicspecifierdesign | september 2011 | 53
powering fpgas
mises its dependency on capacitors, filtering
and EMI shielding.
Lowering height for better air flow
The demand to shrink an FPGA based system’s
size whilst increasing its functionality, memory
storage or computational power, has prompted
designers to refine the techniques used to cool
the components.
One simple method is to provide an efficient
flow of air over the components. Tall components obscure the flow over the thinner packages such as FPGAs or memory ICs. In the case
of pre-fabricated dc/dc point of load regulators, the blockage is severe as
these devices stand some
6~10 times higher
than the FPGA and
other ICs.
The thin BGA
packaging of
FPGAs is extremely helpful in the efficient dissipation of
internally generated heat
from the top of the package.
This benefit is diminished when a taller device
such as a prefabricated dc/dc regulator inhibits
airflow and casts a ‘shadow’ on the neighbouring device.
50% fewer components:
triple & dual output µModule regulators
A new family of dc/dc µModule regulator
systems with multiple outputs and inputs has
been designed to reduce not only component
count but also the cost associated with component insertion on the PCB, whilst eliminating
layout errors and delivering an already made,
complete solution, easing the job of any digital
system designer.
Aside from reducing the dc/dc regulator circuit
materials cost for FPGAs and FPGA based systems by up to 50% and reducing PCB area by
10% compared to a discrete approach, this new
family of multiple output dc/dc µModule regulators shares these benefits:
❏ Current mode architecture
for precise pulse by pulse output load current delivery.
❏ Output current sharing
to increase the output
power of one µModule regulator.
❏ µModule to
µModule device output current sharing to
further increase output power.
❏ Shared input power:
for example, the dual 8A
LTM4616 can be powered from
3.3V and 5V rails simultaneously
to share power from two separate
input supply sources if one input source has insufficient power.
Moreover, this family of dual and triple output
dc/dc µModule regulators addresses the issues
described earlier:
Table 1: Triple & DUal DC/DC µModule Regulators
54 | september 2011 | electronicspecifierdesign
www.electronicspecifier.com
❏ Managing multiple FPGA or system rails.
These dc/dc µModule devices feature tracking and/or sequencing functions for proper system start up and shut down with power sequencing restrictions.
❏ Lowering voltage ripple noise and capaci-
tor requirements.
Each µModule device includes an internal output and input bypass capacitor. In addition,
each device is stable with low ESR output capacitors so the user can adjust the type and
value of the output capacitor to optimise output
ripple or noise.
Figure 1: The triple output LTM4615 is a complete system in a single package, with two outputs of 4A and one of 1.5A, optimised for regulating three voltages from 5V or 3.3V input rails.
❏ Lowering height for better airflow.
At only 2.8mm high, this family of dual and
triple output µModule regulators allows a
smooth airflow for better heat removal of the regulator as well as the components in its vicinity.
One member of the family is LTM4615, a complete triple output dc/dc µModule regulator
system containing two 4A switch mode regulators and one 1.5A VLDO regulator in a
15x15x2.8mm surface mounting package.
Output voltage regulation for each of the two
switching regulators is adjustable over 0.8~5V,
with the third output adjustable over 0.4~2.6V,
making it suitable for powering new generation, lower voltage, multiple rail, very fine
geometry digital ICs.
LTM4619, in the same family, can operate
from input supplies as high as 26.5V. Table 1,
opposite page, lists all the devices in the family
of dual and triple µModule regulators.
Linear Technology’s innovation in dc/dc regulator architecture and packaging has realised
a new generation of point of load solutions
that meet the stricter requirements of FPGA
based systems.
The µModule dc/dc regulator family, as
shown in Table 1, comprises multiple output
www.electronicspecifier.com
Figure 2: The dual output LTM4619 delivers
4A from input supplies as high as 26.5V in a
15x15x2.8mm LGA.
products with a variety of power levels and features. The reliability of these devices is also setting a new standard in multi-chip packaging,
backed up by Linear’s stringent qualification
and testing operations.
These µModule dc/dc solutions are now
paving the way for a new generation of FPGA
and FPGA based systems to boost performance
❚❘
from smaller devices.
Afshin Odabaee is µModule Product Marketing
Manager at Linear Technology.
More from Linear Technology
Return to contents page.
electronicspecifierdesign | september 2011 | 55
Redefining
Industry Expectations
GREEN•POWER
Open-frame & Enclosed
• 5 - 3000 Watts
• PCB or chassis mount
• Ultra-compact design
• Industrial & medical approvals
• Green Power versions
DC-DC Converters
• 0.25 - 600 Watts
• Regulated & unregulated versions
• 2:1 & 4:1 input ranges
• Industry standard SIL, DIL
& SMD packages
Visit our website to request a copy of our
new 2011/12 Power Supply Guide and
see our complete line of power products.
External Power Supplies
• 8-250 Watts
• Energy efficiency level V
• Industrial & medical approvals
• Compact high efficiency design
• Optional class II
DIN Rail Mount
• 5-960 Watts
• Single or 3 phase input versions
• DC OK signal & LED indicator
• Rugged design for industrial
applications