Professional Documents
Culture Documents
00
LIBRARY
FOCUS ON:
AUTOMOTIVE A compendium of articles
from Electronic Design contributors
Sponsored by
INTRODUCTION
Whereas the automobile used to be a mechanical-engineering wonder, the amount of electronics in
today’s automobiles is staggering. As the industry moves toward autonomous vehicles by incorporating
more features of advanced driver assistance systems (ADASs), multiple systems have to work together.
At the same time, carmakers are figuring out how to power electric and other alternative vehicles,
advance charging and battery technologies, and enable infotainment capabilities that bring the office
and living room to the car.
This e-book looks at the design challenges resulting from these trends and delves into how they are
being conquered, as well as the related tradeoffs. It also provides a baseline for how automakers like
Ford plan to incorporate technology and pave the way to autonomous vehicles. For those wondering
what place AI and machine learning will play in this process, we’ve provided a fundamental article
on the types of machine learning we can expect. Enjoy this first overview of the myriad technologies
revolutionizing the modern automobile and stay tuned for further offerings and coverage!
TABLE OF CONTENTS
CHAPTER 1: WHAT’S THE DIFFERENCE BETWEEN MACHINE LEARNING TECHNIQUES?....................2
CHAPTER 2: BMW AND AUDI WANT TO SEPARATE VEHICLE HARDWARE FROM SOFTWARE ...........7
CHAPTER 3: UNDER THE HOOD OF THE EVs IN THE FIA FORMULA E CHAMPIONSHIP.....................9
CHAPTER 4: DUAL-CORE SoC UTILIZES FD-SOI FOR LOW POWER AND HIGH PERFORMANCE........12
CHAPTER 5: LED DRIVERS EXPAND CONTROL OF AUTOMOTIVE LIGHTING...................................15
CHAPTER 6: SNEAK PEEK: FORD SHARES PLAN FOR THE AUTOMOTIVE TECHNOLOGY FUTURE.....17
MORE RESOURCES FROM ELECTRONIC DESIGN................................................. 19,20
1 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY FOCUS ON: AUTOMOTIVE
CHAPTER 1:
A
rtificial intelligence (AI), machine learning (ML), are still far away from achieving what movie producers and sci-fi
and robots are the sights and sounds of science fic- writers have envisioned, the state of AI and ML has progressed sig-
tion books and movies. Isaac Asimov’s Three Laws nificantly. AI software has also been in use for decades but advanc-
of Robotics, first introduced in the 1942 short story
“Runaround,” became the backbone for his novel I, 1. The movie I, Robot has robots that should be following
Robot and its film adaptation (Fig. 1). Although we Asimov’s Three Laws of Robotics.
2 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 1: DIFFERENCE BETWEEN LEARNING TECHNIQUES
es in ML, including the use of deep neural networks (DNNs), are Semi-supervised learning includes a mixture of labeled and
making headlines in application areas like self-driving cars. unlabeled data. This approach is often used when the structure
AI and ML research has been around since before computers even of the data needs to be understood and categorized in addition to
existed. Of course, they made practical creation and applications allowing predictions to be made.
possible. The challenge has always been trying to keep up with the
hype. Usually the programmers were unable to do so—hence, the 2. Rule-based and Decision Tree Systems
many “failures” of AI. Rule-based and decision tree algorithms are the easiest to
In practical terms, AI essentially went underground, providing understand. Rule-based systems consist of a collection of logical
everything from expert systems to behavior-based vacuum rules or conditions based on inputs. A rule is triggered when its
cleaning robots like iRobot’s Roomba. The latter used an 8-bit conditions are met. The triggered rules may change internal state
microprocessor running a behavior-based rule system. Likewise, variables, as well as invoke actions.
e-mail spam filters have been using Bayesian statistical techniques For example, a robot may have a number of sensor inputs that
for decades, with varying levels of success. detect obstacles by touch, as well as inputs about its movement. A
AI is a very large area of research of which machine learning is rule might cause the robot to stop if it is moving and an obstacle
only one part (Fig. 2). The three examined here will be rule-based sensor is triggered.
systems, Bayesian and statistical algorithms, and neural networks. Rules can generate conflicting actions in which case some
These are presented in more detail later. There are more machine priority mechanism needs to be implemented. For example, one
learning approaches not included in this list. rule action may stop a robot, while another wants to change its
direction.
1. Learning Styles A rule- or behavior-based system normally moves from one
In addition to machine learning algorithms, there is the style of state to the next, applying all the rules to each state. Not all rules
machine learning that can be employed. Some algorithms are more need to be examined depending upon the implementation. For
amenable to certain styles that include: example, rules may be grouped by inputs, and some only need to
• Supervised learning be examined if an input changes.
• Unsupervised learning Decision trees are a structure rule-based system where each node
• Semi-supervised learning in a tree has conditions that allow classification by refinement as an
Supervised learning has labeled training data, such as an e-mail algorithm traverses the tree. There are many popular algorithms in
that has been marked as spam. The training process usually this space, including Classification and Regression Tree (CART)
generates improved accuracy over time. It is used in algorithms and Chi-squared Automatic Interaction Detection (CHAID).
like back propagated neural networks. One advantage of these approaches is the ability to backtrack
Unsupervised learning does not have labeled data, and the result the logic process used to perform an action or reach a conclusion.
of the training is often unknown. This approach can be used for This is very valuable for debugging, but it can also be useful in the
creating general rules. learning process. There is also the possibility of examining rules
for proofs.
Finally, a simplistic presentation of a rule-based
Artificial Intelligence (AI)
system where condition A and B invoke action C
overlooks the possibility, and often requirement, that
Machine learning (ML)
the logic involved is binary in nature. Probability
can be used in a rule system, as well.
Neural networks
Rule- and
behavior-based
Bayesian
and Deep
3. Bayesian and Statistics
systems statistical algorithms neural Bayes’ theorem describes the probability of a
networks
(DNN) test result based on prior knowledge of conditions
that might be related to the result. The theorem
relates the chance that an event A happened given
the indicator X, Pr(A|X), to the probability of
event X given A, Pr(X|A). It allows for correction
2. Artificial intelligence is a very large area of research of which machine of measurement errors if the real probabilities
learning is only one part. Three of the major areas of machine learning are are known. Of course, test results come with test
shown here, but there are literally dozens that can be grouped in various ways. probabilities.
There are a number of Bayesian algorithms based
3 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 1: DIFFERENCE BETWEEN LEARNING TECHNIQUES
Neural network Deep neural network configuration options. Of course, the key to a system’s operation is
Input Hidden Output Input Hidden Hidden Hidden Output
the weights associated with the connections and nodes.
Neural networks require training to generate the weights used
in the system. Systems can learn dynamically, but these tend to be
more complex as training normally requires more computational
horsepower than inference. Training uses a feedback system where
an input is matched with outputs and the internal hidden layer
weights are adjusted. This process requires more than just a few
3. A neural network consists of a set of inputs and outputs with samples. A system with a large number of inputs like a photo image
a hidden layer in between. A deep neural network has multiple and a large number of sample sizes will often require a cluster
hidden layers. or high end CPU and GPU to create a DNN configuration. On
the flip side, a microcontroller may often be sufficient for some
on the theorem, including Naive Byes and Bayesian Belief Network applications to utilize a DNN configuration to perform inferences
(BBN). Like differential equations, the theory can be hard to in real time.
understand, but the application is usually straightforward. As FPGAs and specially designed hardware to address neural
noted earlier, Bayes algorithms have been utilized in applications networks are also available. There is even a hardware-based
like e-mail spam filtering, but they are not limited to this narrow network on Intel’s compact 32-bit, Quark-based Curie system-on-
application. chip (see “What Is Inside an IoT Chip?” on electronicdesign.com).
Bayes is only one method that employs probability and statistics. The advantages of this approach are lower power requirements and
There are also regression algorithms that have been used in machine higher performance operation.
learning. Popular regression algorithms include Ordinary Least Google’s Tensor Processing Unit (TPU) is designed to crunch
Squares Regression (OLSR), Multivariate Adaptive Regression 8-bit matrices that are common in DNN computations (Fig.
Splines (MARS), and, of course, linear regression. 4). Unlike number-crunching applications that require double
There are also variants on regression that are used in machine precision floating point, DNN typically has weights that easily fit
learning, such as ridge regression. It is also known as weight decay. into 8-bits. The number of nodes tends to be more important than
It is also known as the Tikhonov-Miller method and the Phillips- weight precision. The TPU will normally be used for training, but
Twomey method. The variants look to simplify the models it could be used in the cloud for running lots of inferences.
to reduce system complexity that provide better generalization DNNs have been used for a wide range of applications, from
support. identifying items in an image to voice recognition. These can be
4. Neural Networks
Artificial neural networks (ANNs) have been around for a long
time, but their high computational requirements for complex
networks has limited the use and experimentation until recently
with multicore systems such as GPGPUs providing an economical
platform for a variant called deep neural networks (DNNs).
Initially neural networks were of interest as a way of copying
biological neural networks like the human brain. The brain is
made up of neurons that are connected via an axon to synapse
and dendrites on other neurons. The electrical signals from
incoming signals are summed by the neuron. A result that exceeds
a threshold sends a signal via the axon.
ANNs are built in a similar fashion, but using electronics or
software. It takes many neurons to perform useful functions;
the human brain has 100 billion. The trick to having something
useful is the way the neurons are connected as well, as the weights
associated with the neurons.
A basic neural network consists of a set of inputs and outputs
with a hidden layer in between (Fig. 3). A DNN has multiple
hidden layers. The networks may be the same logically, but the 4. Google’s Tensor Processing Unit (TPU) has a large, 8-bit
number of inputs, outputs, and hidden layers varies as well as other matrix multiply system designed to accelerate DNN operations.
4 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 1: DIFFERENCE BETWEEN LEARNING TECHNIQUES
ARTIFICIAL INTELLIGENCE USE Neural networks are making many applications practical and
there are a lot of platforms, hardware, and software being brought
Our embedded application(s)
currently incorporate AI - 6% into play (see “A Deeper Look at Deep-Learning Frameworks”
Investigating AI for our
on electronicdesign.com). Electronic Design’s recent Embedded
embedded applications - 32% Revolution survey and white paper highlighted the number of
embedded developers interested in AI (Fig. 5).
No Interest in AI for our
AI and ML have been used productively for decades, even
embedded applications - 62% though AI has been criticized for the hype associated with the
technology. DNNs and CNNs are currently being hyped, but
they too are delivering on the promise. The trick is to understand
that there are many techniques that can be used for a particular
application, and to do more research to find out which will work
VIEWS OF ARTIFICIAL INTELLIGENCE best. n
5 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY FOCUS ON: AUTOMOTIVE
CHAPTER 2:
B
MW and Audi, who often pioneer much that is new standard I/O ports and so on.”
in automotive electronics globally, are in the process Mueller noted that the introduction of Unix more than 40 years
of developing radically new electronics architec- ago in the IT industry allowed data center operators to source
tures for future vehicles. The carmakers are taking their application software from different suppliers. “They were
a similar approach: high-performance central com- no longer stuck with a monolithic supplier responsible for both
puting units replacing today’s outmoded distributed the software and the hardware,” he explained. “Software from
computing architecture. suppliers A and B could run on hardware from suppliers X and
The auto industry is facing profound disruption, with new Y…So in my vision you could think of the car of the future as a
competitors from the IT and consumer electronics worlds data center on wheels.”
making inroads. Google, Apple, Tencent, Uber, Alibaba, and BMW and Audi are developing end-to-end architectures that
Baidu are developing revolutionary new mobility solutions. connect wirelessly from the vehicle’s onboard electronics to the
Tesla is already here. Automated driving is coming faster than carmaker’s back end, and from there to HERE’s back end in the
initially thought. The vehicle’s interior is quickly going digital. cloud, via the internet (see the Audi block diagram at the top of page
A revolutionary new architecture is needed that can take advantage 8). HERE, the digital map company jointly purchased by BMW,
of what has become the state of the art in consumer electronics: Audi and Mercedes-Benz, will use crowdsourced data to provide
internet connectivity, cloud computing, swarm intelligence, and precision maps suitable for autonomous driving and location-
over-the-air feature updates. based services.
Today carmakers purchase infotainment systems and vehicle
control systems from tier-one suppliers who tightly embed Benefits to Carmakers of Separating Software and
software components within electronic control units. If a carmaker Hardware Development
wants to change suppliers, it must validate and test a completely • Flexibility: OEs can more easily switch suppliers and update
different software stack—an enormously time-consuming and vehicles with new features after sale.
expensive endeavor. • Far easier to develop software on known, standard operating
Thomas Mueller, Audi’s new top electrical and electronics systems/hardware platforms
engineer, wants automotive systems to be as mature in terms of • Faster deployment by eliminating redundant validation and
hardware decoupling as IT systems have been for decades. In testing of reused software
a conversation earlier this year, Mueller outlined his vision for • Quicker deployment of diverse implementations of
a new vehicle architecture, which he expects to realize major functionalities—e.g., fault-tolerant versions
elements of by the start of the next decade. “There might be one, • Improved engineering productivity
or a few, central computers,” he said. “We haven’t decided yet. In • Minimized risk of bringing updated features to market
order to decouple hardware from software we need a standard • Proven application software can be reused
operating system like Linux, standard Ethernet connections,
7 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 2: BMW AND AUDI
8 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 2: BMW AND AUDI
AUTHOR BIO:
PAUL HANSEN writes and publishes The
Hansen Report on Automotive Electronics,
a newsletter dedicated to researching and
identifying technology and business trends
in the global automotive electronics industry.
Hansen began writing the newsletter in 1988 as
an outgrowth of his strategy and market research
consulting business, Paul Hansen Associates.
9 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY FOCUS ON: AUTOMOTIVE
CHAPTER 3:
I
n case you haven’t heard of it, there is an electric-vehicle street
racing series called Formula E. The racing season runs from the
fall through summer with a third, season-ending race in New
York (this past July 15). The goal of Formula E is to promote
the concept of sustainability to reduce our footprint as much as
possible while having a positive impact on both people and the
planet. The race features 10 teams with 20 drivers and 40 electric
racing cars, which are powered by lithium-ion batteries. They race
at speeds up to 220 km/h.
Here are some of the most important technologies behind the
Formula E race cars:
Generators: The generators that are used to charge the racing
cars run on glycerine. This carbon-neutral fuel is safe, odorless,
non-toxic, and water soluble. It also is an abundant byproduct
of biodiesel production. A British company called Aquafuel
Research Ltd. was commissioned to build the mobile generators
that could be packed inside a shipping container and freighted
to each race (Fig. 1). The generators are based on standard
production diesel engines—a Cummins KTA50 to be precise—
that have been adapted with Aquafuel’s patented technology to
run on glycerine. The Aquafuel generators on site provide 42 kW
of electricity per car.
Battery: The battery is a sealed black box, meaning that no
access or maintenance by the teams is allowed. For safety reasons,
rather than changing batteries halfway through the race, the
drivers change cars instead.
Williams Advanced Engineering has supplied the batteries (Fig.
2) since the race’s first season. It also was the first company to
complete FIA (Federation International De L’Automobile) crash
testing and has met UN transportation testing regulation 38.3 1. The glycerine base that powers the generators can be made
for the safe air transportation of lithium ion batteries. For Season by salt-water algae where food crops won’t grow.
10 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 3: UNDER THE HOOD OF THE EVS
the final area has been in the detail design around supporting the
cells, because the pack cells of the battery have to be supported and
insulated from vibration and then connected into the bus bars. So,
we’ve made quite a lot of detail changes to do that throughout. All
of those things help us with making the batteries more reliable for
the teams, but they also help us in extracting more power and help
us to put in more power.”
Starting in the 2018 season, however, the battery packs will come
from a new supplier. Actually, McLaren Applied Technologies,
Lucid, and Sony will partner on the battery packs for Formula E
with a contract of two seasons. The new battery technology will
double the current capacity, enabling the drivers to complete a full
2. The Formula E batteries built by Williams for Formula E use race in a single car instead of changing or changing their vehicle
lithium-ion cells made by XALT Energy. mid-race.
The original FIA battery specifications included a 200-kg (441-
3, Williams Advanced Engineering has introduced an updated lb.) cell-weight limit, a 200-kW peak power limit, and a maximum
version of its 28kW. usable energy of 28 kWh. In revised specifications starting with
According to Paul McNamara, technical director at Williams Season 5, cell weight has been nudged to 250 kg (551 lbs.) and
Advanced Engineering, “We have made quite a range of small peak power rises to 250 kW (with usable energy very nearly
changes in three areas: the first one is the cooling system - with doubled at 54 kWh).
changes to make that more robust and make it easier for us Inverters: For the first season, an inverter was supplied by
to assemble and disassemble. We’ve also done a few things to McLaren Applied Technologies. It was mounted above the battery
improve the performance of the thermal interface materials and behind the chassis’ roll hoop. This placement was due to a lack
the details of how the cooling tubes link up with the cells and the of space inside the casing around the eMotor. In this position,
bus bars.” however, the inverter also was accessible and easy to cool.
“The second area is the bus bars themselves,” he continues. “In As part of the Formula E roadmap, which allows manufacturers
the battery, we’ve got proper bus bars dealing with the quite high to develop new technologies, Season 2 opened the regulations to
currents. In terms of building up the battery, those bus bars all allow manufacturers to design and build their own powertrains.
have to be interconnected. They’ve got special joints in them so In season 3, silicon-carbide (SiC) technology is now being
that we can disarm it and bring it down to beneath 50 V and allow used in inverters. For example, Rohm Semiconductor partnered
us to disassemble it. Those joints have improved so they’re more with the Venturi Formula E team to incorporate SiC Schottky
resilient to vibration and the thermal displacement going into Barrier Diodes (SBDs) in the inverter. Their goal was to enhance
them has improved in the battery as well.” McNamara states, “And powertrain performance while increasing switching speeds and
improving thermal efficiency (Fig. 3).
With the Formula E’s success as a technology incubator, it’s
clear that the transportation industry is already benefiting from
technology adoption and development as well as new jobs.
Upcoming seasons will increasingly highlight the potential for
sustainability in EVs. Expect continuous efforts to reduce the
energy footprint while removing barriers to EV adoption, such
as EV infrastructure and the high costs and limited range of
batteries. n
11 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY FOCUS ON: AUTOMOTIVE
CHAPTER 4:
DUAL-CORE SoC
UTILIZES FD-SOI FOR LOW POWER
AND HIGH PERFORMANCE
NXP is employing 28-nm FD-SOI technology with its new i.MX 7ULP family, delivering a high-
performance, low-power application processor.
D
esigners are employing a range of techniques, also makes the gates immune to latch up. The approach provides
like Ambiq Micro’s subthreshold designs (see a 10× to 100× improvement in soft error rate (SER) performance.
“Subthreshold Cortex-M4F Design Sips Less FD-SOI also helps in reducing device parasitics, and the
Power Than Cortex-M3” on electronicdesign. use of device back bias allows performance to be maintained
com) to reduce power consumption while while lowering VDD. There is improved analog peripheral
delivering high-performance devices. NXP is performance with higher gain, better matching, and lower 1/f
employing 28-nm FD-SOI (28FDS) technology with its new i.MX noise enhancements. The gate-first integration removes density
7ULP family, targeting applications ranging from wearables to rules for precision analog. Device tuning with back biasing is
automotive. The i.MX 7ULP family is the low power migration employed to compensate for process variation.
path (Fig. 1) from the popular i.MX6 family. NXP developers used domain and subsystem optimizations to
FD-SOI utilizes an ultra-thin buried oxide layer (Fig. 2) to design the chips, along with a custom standard cell library with
improve electrostatics, thereby enabling shorter gate lengths. It mixed voltages. They employed a large dynamic gate and body
biasing voltage (VBB) range to address speed and
power requirement tradeoffs and to deliver an
i.MX Processor Roadmap
extreme low leakage implementation.
The system actually use a dual-core, asymmetric
i.MX 6QuadPlus i.MX 8 family
ARM v8-A design with an ARM Cortex-A7 paired with an
i.MX 6Quad (32-bit/64-bit) Advanced graphics and performance
ARM Cortex-M4 (Fig. 3). The cores can operate
i.MX 8M family
i.MX 6DualPlus
Advanced audio and video
independently when it comes to operation and
i.MX 6Dual sleep mode. Versions of the i.MX 7ULP are
i.MX 8X family
i.MX 6DualLite Safety-certifiable and efficient performance
available with the 3D Vivante GC7000 NanoUltra
3D GPU with a low-power single shader, or the
i.MX 6Solo
GC320 Composition Processing Core (CPC) for
i.MX 6SoloX 2D graphics support.
i.MX 7 family
i.MX 6SoloLite Flexible efficient connectivity The chip has secure boot support for both
i.MX 6UltraLite
ARM v7-A
i.MX 7ULP family
cores and the QPSI interface supports on-the-fly
(32-bit) encryption and eXecute-in-place (XIP) support.
Ultra-low power with graphics
i.MX 6ULL
The Cortex-M4 has access to the typical serial and
ARM v7-A digital peripherals as well as a set of dual-channel
1. NXP’s 28-nm FD-SOI technology is used to build the i.MX 7ULP family that 12-bit ADCs and DACs. The Cortex-A7 sports
replaces the low end of the popular i.MX6 family. USB 2.0 host and OTG interfaces. The TrustZone
12 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 4: POWER SUPPLY FUNDAMENTALS
Gate Gate
to view this article online, ☞click here
Source
e Drain Source e Drain ☞ BACK TO TABLE OF CONTENTS
Ultra-thin buried oxide
Gate 0 1.3V
Source Drain
Ultra-thin buried oxide
3. The i.MX 7ULP uses a dual-core, asymmetric design with an ARM Cortex-A7
paired with an ARM Cortex-M4.
13 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY FOCUS ON: AUTOMOTIVE
CHAPTER 5:
LED DRIVERS
EXPAND CONTROL OF
AUTOMOTIVE LIGHTING
LEDs in automotive interior and exterior lighting devices that implement a simplified driver IC
design can have their brightness intensity adjusted depending on the scenario.
T
he automotive industry makes extensive use of are the fundamental design parameters for the correct driver.
light-emitting diodes (LEDs) in high-beam head- Fluctuations that occur with the vehicle’s power supply must
lights, brake and position lights, and side and rear therefore be strongly considered. Moreover, other requirements,
direction lights. In an LED driver design, the bright- which are reflected in the design phase, include temperature
ness intensity of the various illuminating devices is and humidity, voltage range, electromagnetic interference (EMI)
not equal; rather, it depends on their specific func- and compatibility (EMC), as well as the reliability requirements
tion. Needless to say, then, that LEDs operate at different bright- dictated by the qualification tests.
ness levels—for example, at full brightness for braking and from High-reliability demands in automotive applications indicates
10% to 25% for the rear lights. LED dimming circuits are used to that protection circuitry is essential within the IC driver, which
differentiate the brightness level through a
pulse-width-modulation (PWM) driving T T T
technique, which modulates the width of
the current pulses applied to the LEDs. LED
driver solutions integrate a PWM system to V
DIM
control the brightness by providing a ramp
generator, thus simplifying the driver design. D DMIN DMAX
PWM has been adopted as the preferred
dimming technique for high-quality LED tD tSU tSD tD tSU tSD tD tSU tSD
lighting. An essential aspect of the lighting
control system is the power management IF
provided by integrated-circuit (IC) drivers
in several configurations, such as buck
and buck-boost topologies. The ease of
1 tD + tSU T – tSD
controlling LEDs makes them an intelligent T= DMIN = DMAX =
fDIM T T
lighting system.
1. The delay exhibited by the LED driver is in response to its PWM signal. These
LED Driver Features delays result in the contrast ratio (CR) factor of the system. Here, tD represents
LED solutions require a constant current the propagation delay from when VDIM (PWM signal) goes high to when there’s
in order to produce uniform brightness. The a response from the If (forward current) driving the LED (tSU and tSD are the LED
accuracy of the source and the fluctuations forward-current slew-up time and slew-down time, respectively). DMIN and DMAX are
in terms of voltage and other parameters the minimum and maximum duty cycle, respectively.
14 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 5: LED DRIVERS
Driver Topologies +
LEDs typically require constant current to produce a uniform
light output. Therefore, an LED driver must be able to vary the 3. The inductor, which is discharged through the diode,
output voltage to maintain a constant current. The output voltage provides current to the load in this buck-boost converter.
is related to many parameters, such as the temperature of the LED
matrix and the number of LEDs in series. The designer must be L1 CS S2
able to predict with great accuracy the maximum output voltage in
+
order to select the optimal regulator topology and, therefore, the +
+
driver IC and associated components. VIN S1 L2 C R VOUT
thereby prolonging the life of LEDs. The quality of the light 4. A SEPIC is similar to a traditional buck-boost converter, but
produced is determined primarily by the light-intensity stability, has the advantage of a non-inverted output (the output voltage
which requires a precise regulation of the current with constant has the same polarity of the input voltage) and isolation
working points in all voltage and temperature conditions. It between input and output (provided by a capacitor in series).
15 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 5: LED DRIVERS
Depending on whether the LED application is automotive duration LEDs—high performance, high brightness efficiency,
or general lighting, use of a multi-topology LED driver with and low-heat-dissipation energy consumption—are the main
maximum flexibility of input- and output-voltage range makes features of an ideal automotive lighting solution.
it easier to select the correct driver. The voltage range for the As more lighting technologies come onto the scene, new LED
automotive sector extends from 9 to 16 V (nominal 14 V) and driver technologies for internal and external lighting help produce
includes extreme conditions, such as reversal of the battery an added level comfort in a wide range of vehicles. The goal is
polarity (‒12 V), fault conditions like load dump (which occurs to provide linear LED dimming with a large contrast ratio. The
when the battery is disconnected from the alternator), and other correct procedure is to operate the LEDs at the manufacturer-
transient voltages. recommended forward current/forward voltage.
Among their other benefits, the buck-boost SEPIC configurations
ensure constant brightness in all battery-voltage variations. When Conclusion
the need arises to control several LEDs in series, a buck-boost LEDs are becoming a significant force in the lighting market
topology can address a variety of application requirements, due to their long lifetime and the ability to control specific lighting
including the ability to manage extreme voltage values. requirements. More integration of systems-on-a-chip (SoCs) will
In some automotive exterior-lighting applications, the LED array continue to reduce the size of driver ICs, leading to faster product
or matrix may be located at a distance from the driver/controller. development cycles and accurate management of high-level
In these cases, a boost converter can be a more appropriate lighting features.
topology choice. Carefully choosing the buck regulator can allow The LED driver market for lighting is estimated to have a
for PWM dimming frequencies in the kilohertz range. While this compound annual growth rate (CAGR) of about 27% in the near
feature perhaps isn’t necessary for traditional lighting, it can be future. The key driving factors responsible for the upswing in the
effective in applications such as high-speed stroboscopic effect for LED driver market includes the greater efficiency exhibited by
recognition activities (imaging) in the industrial and automotive power-management circuits and the strong demand for LEDs in
sectors. commercial and industrial applications. n
16 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY FOCUS ON: AUTOMOTIVE
CHAPTER 6:
SNEAK PEEK:
FORD SHARES PLAN
FOR THE AUTOMOTIVE
TECHNOLOGY FUTURE
At the recent Manufacturing & Technology Expo, Ford’s VP of Research & Advanced Engineering
provided an inside look into how the company plans to leverage the avalanche of technology
advances.
C
armakers have a host of new technologies avail- as the company’s inspiration today, saying “Henry Ford put the
able now, and a seemingly limitless amount of world on wheels.” Indeed, Ford was the first to provide a $5/
options are evolving to provide future options. day wage in the United States, creating a class of workers who
Most people assume that these options will be could afford to buy the product they made. The company plans
limited to the high-end luxury-vehicle market or to continue to be a company of “firsts” that are needed by society
cutting-edge vehicles like the Tesla line. However, and in the car industry by focusing its efforts on the following
Ford’s SYNC—now in at least its
third iteration—has already high-
lighted the company’s efforts to
provide ease of use with technol-
ogy applications and tools early
on, ranging from running a linked
cellular device through the cen-
ter-console display to voice-rec-
ognition systems for maps, etc. In
a keynote address at the Manu-
facturing & Technology Expo, Ken
Washington, vice president, Re-
search and Advanced Engineering
at Ford Motor Co., touched on
how Ford plans to similarly make
available better vehicles and a ve-
hicle experience via data analytics,
the Internet of Things, virtual re-
ality, robotics, and additive manu- 1. Talking about Ford’s aggressive work on autonomous vehicles, Ken Washington, vice
facturing (Fig. 1). president, Research and Advanced Engineering at Ford Motor Co., notes that most people
Washington credited the are surprised to learn that Navigant Research named Ford the No. 1 car manufacturer in
original innovative Henry Ford automated driving systems.
17 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
ELECTRONIC DESIGN LIBRARY CHAPTER 6: FORD SHARES PLAN
also leading to emerging mobility services and solutions. ☞ BACK TO TABLE OF CONTENTS
4. Autonomous vehicles: Washington sees vehicle autonomy
as “as important a shift in the automotive industry as the moving
assembly line.” At Level 4 autonomy, no driver is required within
a defined geographic area. The deployment of such vehicles
18 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
GET MORE IDEAS AND SOLUTIONS FROM
ELECTRONIC From the start of your design
idea, sketched as a schematic
on a napkin, through to the
DESIGN NOW! ordering process for prototypes
and final production …
19 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics
CHECK OUT THESE RESOURCES FROM
ELECTRONIC DESIGN AND OUR SISTER BRANDS
WEBSITES
MAGAZINES NEWSLETTERS
You can also apply for a subscription to one of our Stay current with the industry with newsletters
traditional magazines available in both print and created by engineers and editors with inside
digital formats. connections! You’ll find coverage on a wide variety
of electronics technologies from Analog & Mixed
ELECTRONIC DESIGN – MICROWAVES & RF –
Signal to Embedded and more.
complimentary in US and complimentary internationally-
Canada - ☞ Subscribe Now ☞ Subscribe Now ☞ Click Here to check out what more than
Non-qualified or Outside the Non-qualified: you will be 200,000 engineers are reading now.
US and Canada: you will be prompted to pay based on your
prompted to pay based on your location.
location.
THE SMART HOME BUTLER
The Real Smart IoT
Electronic Design
the authority on
emerging technologies march 2017 trusted eNgiNeeriNg resource for oVer 50 years www.mwrf.com
for design solutions
Impedance Tuners
APRIL 2017 electronicdesign.com
in control 100
build, buy, p|
or both?
consumer electronics
•
$10 computing Price Point
•
trends in Poe
Y
march 2017 • Vol. 56 •, No. 3
FAST
CM
MY
CY
APRIL 2017 • Vol. 62, no. 11
CMY
$10.00 Periodicals Postage Paid • USPS 100 Approved Poly $10.00 Powered by Periodicals Postage Paid • USPS 100 Approved Poly
ABOUT US
A trusted industry resource for more than 50 years, the Penton Electronics Group is the electronic
design engineer’s source for design ideas and solutions, new technology information and engineering Powered by
essentials. Individual brands in the group include Electronic Design, Microwaves & RF and Power
Electronics. Also included in the group is a data product for engineers, SourceToday.com.
20 | electronicdesign.com ☞ Want more coverage like this? Subscribe to Electronic Design’s newsletter Automotive Electronics