•  
• EDN Europe MAGAZINE
•  
• Free SUBSCRIPTION
• Subscription Renewal
• View Media Kit

• EDN
• EDN Australia
• EDN Asia
• EDN China
• EDN Japan

Silicon safeguards automotive circuits

TRANSIENT VOLTAGES ARE AN INEVITABLE FEATURE OF THE AUTOMOTIVE ENVIRONMENT AND CAN EASILY DAMAGE DELICATE SILICON. INCREASINGLY HOWEVER, AUTOMOTIVESTRENGTH SILICON SOLUTIONS ASSUME ROLES THAT PASSIVE COMPONENTS TRADITIONALLY TACKLE.

In response to legislation that deals with environmental and safety issues together with consumers’ insatiable desire for creature comforts, the automotive- electronics sector continues its healthy growth pattern. In Europe alone, automotive-semiconductor sales totalled about $7 billion in 2005, grew at around 12% in 2006, and now account for 32% of all semiconductor sales in Germany—the largest automotive market in Europe and its most significant electronics sector. Meantime, developments continue: for example, car makers in the US prepare for legislation that mandates electronic stability-control systems in every car and small truck as of the 2012 model year—a move that the country’s National Highway Traffic Safety Administration estimateswill annually save between 5000 and 10,000 lives. Researchers consider that this little-promoted technologyhas life-saving potential that’s secondonly to that of seat belts, with innovationssuch as airbags and anti-lockbrakes coming nowhere near close.

Whatever the target application, protecting delicate semiconductors from the rigours of the automotive environment is crucial to any system’s success. As a result, there’s a bewildering range of protection devices that designers can consider. But before discussing the characteristics of any device, it’s worth reviewing the conditions that complicate automotive-circuit protection, from ESD to load-dump events—conditions that typify low-voltage dc applications where electronics meet inductive loads. Unsurprisingly, working through standards that generically fall under EMC regulations is highly complex and the domain of test professionals, but there are several resources that can help simplify the process if you have the need for a formal approach (see sidebar “A quick guide to auto EMC standards”). A practical approach guards against gross overcurrent faults at the primary— distribution-system—level before individually protecting secondary-level subassemblies from dangerous conditions that appear directly on the supply lines or locally couple into your circuitry—leading to the need for multiple protectionelements throughout the vehicle.

KNOW YOUR FUSE

AT A GLANCE

“Simple” fuses have widely varying characteristics. Polymeric devices don’t suit repetitive-fault conditions. Metal-oxide varistors can handle most load-dump transients. Smart switches include comprehensive protection. Overvoltage-protection ICs handle arbitrary power levels.

Virtually every automotive system includes traditional wire-link fuses, most often as a last-resort defence against gross overloads such as wiring-loom failures. The familiar blade-style fuses that appear in most vehicles’ fuse boxes are comparatively large devices that typically have 32V ratings and span current levels from 1 to 60A. For instance, Littelfuse’s ATO series, which appears in countless vehicles, measures 19.1x13.7x5 mm for the fuse alone. Better suiting portable or space-constrained equipment, the company’s right-angle PCB-mounting miniblade fuse and holder measures around 17.5x20x7.62 mm, and an end-stackable, vertically mounting version is available to suit applications such as fuse arrays. While fuse-current ratings span 2 to 30A, be sure to check the fuse holder’s capability as this parameter typically limits application to 15 to 20A. These mini-blade fuses are “fast-acting”, withstanding 110% of rating for at least 100 hours while blowing within 150 msec to 5 sec when subjectedto a 100% overload; the time-to-blow falls to between 30 and 100 msec at 600% ofampere rating (Figure 1).

Selecting a fuse is more complex than this figure suggests: the protection method depends on melting a metal element under conditions that depend upon the amount of energy (which depends on the square of the current I) that flows through the element for a given time period, t. Littelfuse characterises this I²t term by applying a current pulse of just sufficient magnitude to melt the element-under-test within 8 msec (or within 1 msec for thin-film devices), with 10 sec between pulses. This ensures that all of the I²t pulse energy goes into melting the element; thereafter, a short arcing period exists prior to the fuse opening to interrupt current flow. Importantly, the I²t parameter remains constant for a particular fuse design, independently of ambient temperature or test voltage. As a selection parameter, this characteristic is most useful in determining a fuse’s ability to withstand transients by calculating the amount of pulse energy that you need to guard against, including any inrush-current peaks that your circuit may generate. Based on experimental results, a paper from the Society of Automotive Engineers examines the effects of transient overcurrent pulses on the lifetime of automotive blade fuses with the objective of properly selecting such devices(Reference 1).

In general, blade and other traditional fuses last indefinitely providing that you operate them at no more than 75% of their steady-state current rating. Because vendors typically specify parameters at +25°C ambient, don’t forget that fuses with a conventional construction—that is, fast-/very fastacting wire links and wire-wound slowblow types—require derating to around 90% of their nominal-current rating at +150°C. But if the fuse uses a dualelement slow-blow anti-surge or thinfilm construction, the derating factor can be dramatically larger, falling to maybe 65% of nominal at +100°C andas little as 25 to 30% at +150°C.

The cold resistance of a 20A fuse is typically around 3 to 5 mΩ, so there will always be some self-heating and a current-dependent voltage drop thatyou may need to consider. This consideration equally applies to any seriesconnectedprotection device, most ofwhich constrain losses to under 200 mV.Other key considerations include interruptcapability and the level of voltageand current that a fuse can safely disconnect.Most blade-format fuses have amaximum interrupt capability of 1000Aat 32V, which adequately protectscircuits from the maximum availableshort-circuit current that conventional12V batteries deliver. Suiting emerging42V systems, Littelfuse offers a range offuses that are broadly similar to its 12Vtypes but that offer a 1000A interruptrating at 58V dc.

PTC FUSES IMPROVE SERVICE

Resettable PTC (positive-temperaturecoefficient) fuses from vendors such as Bourns and Raychem are now very popular for protecting circuits at board level. Appearing under the brand names Multifuse and PolySwitch, respectively, these devices depend on the phasechange effect in a crystalline organic polymeric material that often employs dispersed carbon-black particles. In normal operation, these particles offer a low-resistance path that sharply transitions into a high-impedance state when sufficient heat is present. Because the operating principle relies on the balance between I²R power losses in the device and the heat loss into the surroundingenvironment, it’s important to rate devices over the temperature range ofthe equipment you’re trying to protect.For instance, a device that’s rated for 1Aholding current at +25°C may withstandaround 1.5A at -40°C but only 0.5A at+80°C. As a result, Raychem offers arange of automotive-qualified parts thatthe company specifies over ambienttemperatures from -40 to +125°C. Therange spans nominal levels from 0.22Aat 60V to 15A at 16V and is availablein through-hole-mount-disc and -platepackages together with a range of surfacemountoptions.

Typically, a 2:1 relationship exists between a device’s trip current and its maximum normal holding current at room temperature. As in the case of a conventional fuse, the time-to-trip depends upon the severity of the overload. If a device trips due to an overcurrent condition that’s then removed, the device eventually cools down sufficiently to conduct with a resistance of or very close to its pre-trip state. This does not mean that these devices suit repetitive operation in their tripped state—repetitive thermal shock results in the material becoming more conductive, degrading its effectiveness. As one indicator, Raychem’s data shows that the effect of cycling its surface-mount ASMD and AHS devices for 20 cycles of -40 to +85°C reduces the nominal on-resistance values by typically 33%. It’s also important to consider the device’s ability to withstand soldering processes, for which IEC-STD 68-2-20demands 10 sec at +260°C ±5°C.

Favourites within the aerospace industry, magnetic and thermal circuit breakers from brands such as Eaton, E-T-A, and Klixon (now under Sensata’s ownership) withstand repetitive tripping and can also serve double-duty as pushpull switches. Replacing fuses, these traditional circuit breakers frequently appear on aircraft distribution panels, where visual feedback makes it easy to determine their on/off states. These advantages have made such circuit breakers popular in motorsport applications, where reliability and compact dimensions are crucial. For less demanding applications, Sensata now offers the First Technology line of thermal circuit breakers that replace normal blade-typefuses. Notably, the bimetallic-strip-based Maxi Breaker series is available in5A increments from 5 to 50A with a-40 to +125°C ambient-temperaturerating. You can choose between cyclingdevices that automatically resetwhen the overcurrent condition is nolonger present and a non-cycling versionthat requires power removal beforeit resets.

QUASH TRANSIENTS AT SOURCE

From low-energy ESD strikes that last tens of nanoseconds to high-energy surges that systems must endure for hundreds of milliseconds, transients are an inevitable feature of automotive systems. Datasheets for typical surge suppressors such as a metal-oxide-varistor (MOV) or an avalanche diode almost invariably describe a transient whose voltage waveform rises exponentially for 1.2 µsec before decaying over 50 µsec, with a current waveform that follows a similar 8/20-µsec profile. This template derives from lightning-strike analysis and is inadequate for automotive use, where fault conditions other than ESD strikes can be relatively long in duration (Table 1). Because the characteristics of automotive systems vary so widely, any survey of representative fault conditions returns significantly different values— hence you should only ever treat such parameters as a starting point from which to develop protection systemsthat suit your target environment.

Table 1 suggests that systems must withstand reverse- and double-voltage conditions with virtually limitless energy potential. If you set aside issues concerning reverse-voltage protection, voltage clamps that provide primary protection in 12V systems—for instance, in a MOSFET-based equivalent of a traditional fuse box—must not clamp below about 27V, with a typical maximum clamping voltage of around 40V. The worst-case waveform that the protector then has to endure is the so-called load-dump event, which occurs when the alternator loses its connection to the battery while the engine is running. This typically occurs due to corrosion or loose terminals compromising systemconnections.

Because the alternator’s control loop is relatively slow, losing the battery supply injects a positive-going transient into the system that can exceed 100V from a source impedance of maybe 0.5 to 4Ω. This transient typically has a rise time of 5 to 10 msec, decaying exponentially over another 40 to 400 msec. The template that ISO 7637-2 specifies is more benign, rising to 60V in 1 to 10 msec and lasting only 150 to 180 msec with a repetition rate of no more than 30 sec (Figure 2). For this reason, many vendors choose to qualify their systems using more stringent parameters, applying the test pulse to any circuit that connects to the battery either directly or via a switch. The characteristics are broadly similar for 24V commercialvehicle systems except that the peak value is proportionately higher. Notice that while load-dump events are relatively rare, the alternator’s field decays every time you turn off the engine, injecting low-energy negative-going pulses of some -40 to -100V into thesupply line.

Selecting a primary-circuit suppression device heavily depends on the characteristics of the individual system,notably maximum alternator-output power. For instance, a 100A alternatortheoretically outputs some 1500W, soa suppressor with a 40V peak clampinglevel must absorb ≥37.5A. If weassume a load-dump decay period ofabout 150 msec, the energy to dissipatelies in the 80J region, requiring a 100Jdevice for safety. Under worst-caseconditions, there’s always the additionalrisk of arcing between loose orcorroded terminals, but such arcs typicallyextinguish relatively quickly in12/24V systems. This is not the case atthe 42V level, when the higher voltagecan indefinitely sustain arcing, potentiallycreating a fire hazard that systemdesigners must prevent—for example,by replacing mechanical relays withsolid-state equivalents.

Although similar to zener diodes, avalanche diodes for transient-protection applications have wider junctions that permit them to dissipate fault energy more effectively. Transient-voltage-suppression devices that can handle load-dump power levels are available from vendors such as ON Semiconductor, Semtech, Sensonetics, and Vishay. I use Vishay’s SM8A27 as the primary overvoltage suppressor in the powerbox (electronic fuse-box) for a front-running manufacturer’s World Rally Car project. This powerbox distributes over 200A and supplies vicious loads that include massive cooling fans, multiple high-pressure fuel pumps, and the turbocharger’s exhaust-gas-recirculation valve. Available in a surface-mount DO-218AB package with an overall footprint of about 10.5x16 mm, the unidirectional SM8A27 meets the ISO 7637-2 surge specification, withstanding as much as 130A of peak reverse surge current for a 10 µsec/10 msec rising/exponentially decaying waveform—equivalent to 5200W of peak pulse-power dissipation. Its nominal breakdown voltage is 27V with a maximum 40V clamping voltage for a 75A surge. Easily absorbing negative- going transients, the device has an instantaneous forward voltage of under 1V at 100A and withstands as much as 700A for a single 8.3-msec half sine-wave. Leakage currents at the stand-off voltage rating are negligibly small—around 1 µA at room temperature, rising to a maximum of 150 µA at the +175°C maximum casetemperature.

The similar SM8S series suits operation from 10 to 43V, with A-suffix devices, reducing the normal maximum breakdown voltage from 36% above its nominal value, to just 23% above. Connecting a pair of these devices back-to-back provides bidirectional protection with characteristics that you can adjust by selecting appropriate device ratings. Alternatively, MOVs are inherently and symmetrically bidirectional. Typically suiting secondary, board-level protection applications in today’s vehicles, most such devices from vendors such as AVX Kyocera, Epcos, and Littelfuse come in surface-mount packages up to 2220 format with maximum power ratings of 25 to 35J. Alternatively, disk-format MOVs such as the S20K14AUTO from Epcos’ automotive family are available to absorb energy levels up to 100J on both 12 and 24V systems. The company also publishes a mass of application information, including construction details for a hardware load-dump simulator (Reference 2). If you prefer to simulate charging-system faults rather than risking components, Epcos provides PSpice models for these parts. Alternatively, consider Ansoft’s RMxprt/Simplorer suite. Using Maxwell 2D/3D equivalent-circuit techniques, the software allows you to generate static, transient, and harmonic modelsto simulate charging-system events and to observe the effect of adding protectioncomponents.

A QUICK GUIDE TO AUTO EMC STANDARDS

Multiple standards are in use within Europe and North America that help assure the reliability of a myriad of automotive circuit- protection elements. Many of these have their roots in military circles, notably including the US Department of Defense’s MIL-STD-202 test standard method for electronic and electrical parts and the partner MIL-STD-883 for microcircuits— both of which you can download for free from the Defense Supply Center Columbus website. Focusing on automotive-supplier qualifi cation, the Automotive Electronics Council—a cooperative venture that Chrysler, Delco Electronics, and Ford originally formed— resulted in the AEC-Q100 stress-test qualifi cation for ICs back in 1994. Since that time, the now much-expanded group has published AEC-Q101 for discrete components and AEC-Q200, which applies to passives. Many component vendors qualify their parts to meet the requirements of these documents. Newly published in 2007, ISO 7637-3 is the latest addition to the European normative EMC specifi cation that applies to conducted electrical transients in automotive systems (Reference A). Part 1 of the series is the introductory document that defi nes the terms relating to electrical disturbances from conduction and coupling, while Part 2 describes bench tests for injecting and measuring transients along the vehicle’s supply lines. Part 3 tackles transient transmission by capacitive and inductive coupling via lines other than supply lines, including sources such as inductive loads switching and relay contact bounce. The specifi cations also provide a methodology for failure-mode severity classifi cation and cost CHF (Swiss Francs) 48, 132, and 108, respectively. Available for free yet one of the most useful documents available, Ford’s own EMC standards are the baseline for many vendors’ processes (Reference B). The document lists the world’s major applicable standards together with Ford’s comprehensive test methodologies. The company’s EMC Online Web site also presents supporting information such as a worldwide list of approved laboratories and their capabilities, together with yet more invaluable information such as Ford’s EMC design guide for PCBs. More useful resources appear at the Automotive EMC Network’s Web site, a UK-based non-profi t organisation that comprises a group of industry professionals who make automotive EMC information freely available via the Web. As well as listing applicable worldwide standards and describing its own generic EMC standard, the site carries a wide range of papers together with links to equipment and service suppliers. Subscribe to the group’s free newsletter service to keep upto- date with industry news, revisions to standards, events, and other information. REFERENCES ISO 7637-3:2007, “Road vehicles—Electrical disturbances from conduction and coupling, Part 3: Electrical transient transmission by capacitive and inductive coupling via lines other than supply lines,” International Organization for Standardization, www.iso.org. “Component and subsystem electromagnetic compatibility—Worldwide requirements and test procedures,” specification no. ES-XW7T-1A278-AC, Ford Motor Company, October 10, 2003, updated June 7, 2006, http://www.fordemc. com/docs/requirements. htm.

 

SWITCHES PROTECT LOADS

Another method of circuit protection takes advantage of smart load switches, many of which specifically target automotive applications. Representative vendors include Infineon, International Rectifier, NXP, and STMicroelectronics. Essentially, smart-switch devices are N-channel MOSFETs that integrate the onboard charge pump that’s necessary to enhance the high-side power switch together with various protection features. The baseline feature set comprises overcurrent and overtemperature protection, but most devices offer additional benefits that can radicallysimplify system design.

Still the most powerful such device available, Infineon’s automotive-qualified BTS555 comes in a five-lead throughhole- mount TO-218 package that houses chip-on-chip technology (Figure 3). At its maximum +150°C junction temperature, this power switch achieves a maximum on-resistance of just 4 mΩ at 120A. Overtemperature protection cuts in beyond this point, with +10°C of hysteresis between automatic reset attempts. In normal operation, keep the case temperature below +85°C when the device’s nominal load current is a massive 165A. Within case-temperature ranges from -40 to +150°C, short-circuit current-limit values span 200 to 650A, with a maximum timeto- trip of 300 µsec. An overvoltage clamp limits positive-going transients of up to 80V to a 47V maximum, while characteristics that virtually mirror normal positive-quadrant operation clamp negative-going transients. As its datasheet describes, a few external components increase the part’s ability to tolerate conditions such as reversebatteryconnections.

These features suit the device to primary-circuit duties in applications such as electronic fuse boxes. I have used these and the similar, 70A-rated BTS650P in a 32-channel powerbox application without seeing a single failure in the vehicle fleet over more than five seasons of rallying at the sport’s highest level. Other invaluable device features include a current-sense output that mirrors a small proportion of the load current. For the BTS555, this proportion lies around the area of1:30,000, but values between 1:23,000 and 1:61,000 are possible under currentand temperature extremes. Nonetheless,a simple resistor to ground scalesthe current-to-voltage conversion,which you can use to limit the device’soutput current way below its capabilityusing external hardware and/or apply toan ADC to monitor a circuit’s currentdemands. Compensating the currentsensecircuitry using, for instance, alook-up table can substantially improvemeasurement accuracy. Crucially, overspecifyingthe device’s current capabilitywhile externally limiting it substantiallyreduces the requirement for heatsinks,guaranteeing robustness at substantiallyless cost in overall system bulk.

Smart-power technology also appears as an attribute of lower-power multiplechannel subsystem components, such as Freescale’s recent MC33879. One of a family of devices that interface microcontrollers with various inductive, incandescent, and LED loads, this octal switch combines CMOS logic, bipolar/ MOS circuitry, and DMOS power FETs that have a typical on-resistance of 0.75Ω at room temperature. You can individually connect the device’s outputs in high-side or low-side configurations and freely parallel outputs for more current drive. Every output features independent voltage clamping and current limiting from 0.6 to 1.2A, with high-side operation limiting negative- going inductive transients to -20V and low-side operation limiting them to +45V. A 16-bit, 3.3/5V-compatible SPI port allows you to switch outputs and to read back status information. It’s also easy to daisy-chain multiple devices using the serial connection and to control them synchronously via a common chip-select signal from an I/O port. The chip’s protection mechanisms comprise individual overtemperature and auto-retry overcurrent trips for each output, together with an overall device shutdown in the presence of supply- line voltages beyond the chip’s 5.5 to 26.5V range. The chip also features internal reverse-voltage protection on its power-supply input and can tolerate losing its power or ground connection without damaging itself or energisingload circuits.

ICS CONTROL OVERVOLTAGES

A growing number of power-management ICs are available that allow you to precisely tailor overvoltage protection at arbitrary power levels. The general operating principle relies on the chip driving an external series MOSFET, monitoring the supply line and adjusting the gate drive to either clamp or disconnect the power element in the presence of overvoltages. Assuming that the chip has sufficient gate-drive power, you can then size the MOSFET to suit load-current demands. One point that’s potentially a concern is the threshold area between clamping and off states, as the clamp state operates the MOSFET in its linear region, dissipating heat that you must limit tosafe values.

One example of a device that works in this way is Maxim’s MAX6398, which specifically suits load-dump transient control. The chip’s chargepump provides around 75 µA of drive current at levels ≥8.5V above the normal 14.4V rail. In conjunction withthe gate-driver’s risetime of 1 msec into a 6000 pF load, these characteristicswill adequately enhance virtuallyany high-power n-channel MOSFET.Connecting a resistive divider betweenthe input voltage—that is, before theMOSFET—and the chip’s SET pinconfigures the device in overvoltage-switch mode, disconnecting theload from the input at a level that thedivider determines. Alternatively, connecting the divider tothe MOSFET’s outputconfigures the chip as avoltage limiter, modulatingthe gate-drive toregulate the load voltage.The traces in Figure 4 show the MAX6398limiting a load-dumptransient to 20V andthe instantaneous effectof this transient on theoutputs of a downstreamMAX5073 switch-modeconverter.

The MAX6398 operates from 5.5 to 72V with an undervoltage lockout at 5V. The similar MAX6397 includes an always-on, low-dropoutvoltage linear regulator with a choice of 1.8, 2.5, 3.3, and 5V outputs at up to 100 mA. Available in thermallyenhanced 3x3-mm leadless packages, these chips operate at junction temperatures from -40 to +125°C with thermal shutdown activating at +150°C. Similar parts in the MAX6495-99 series provide a range of additional capabilities, including latching the power MOSFET off after an overvoltage event, support for an external P-channel device to provide reverse-voltage protection, adjustable undervoltage threshold, and a power- OK status indicator. Alternatively, consider the MAX16013/4, which drives two P-channel devices in series to combine overvoltage and reversevoltage protection. Evaluation boardsare available for many of these chips.

Also available on an evaluation board, Linear Technology’s new LT4356 reacts to the severity of load-dump and other overvoltages by acting as a voltage limiter or as a disconnection switch by controlling a single external MOSFET (Figure 5). Alternatively, the device will drive a pair of back-toback N-channel MOSFETs to protect the load from reverse-voltage connections. Also optionally, the device provides inrush-current protection by monitoring a maximum 50-mV drop across a low-value series resistor. This feature is particularly useful when the downstream circuitry includes bulksmoothing capacitance.

A capacitor between its fault-timer input and ground sets the period between the device limiting an overvoltage event and disconnecting the load if the event persists; in this case, the FLT pin goes low ahead of shutdown to warn the system of impending power loss. The current source that controls this timer’s period has an output that’s linearly proportional to input overvoltage, helping to keep the power element within its safe area of operation. Accordingly, it’s important to set the timer to ignore short-duration overvoltages while shutting down if the condition is likely to compromise the MOSFET’s safety. To allow adequate advance warning of an impending supply-voltage disconnection, the current source switches from its overvoltage-dependent value to a fixed 5-µA level when the voltage at its TMR pin reaches 1.25V. Assuming that the fault condition persists, the capacitor then continues to charge to 1.35V, at which point the chip switches theMOSFET off.

To maintain IC operation during cold-start cranking and to minimise its own protection needs, the LT4356 has an exceptionally wide operating voltage range of 4 to 80V with a 100V absolute maximum value. It also tolerates reverse voltages to -30V. Other features include a spare operational amplifier whose negative input connects to the chip’s internal 1.25V reference. You can use this amplifier to drive an external PNP transistor to create a linear voltage regulator, or as a comparator that monitors a system voltage level. The LT4356 is available now in a 10-pin MSOP and a 12-terminal leadless DFN package that shrinks itsfootprint to just 3x4 mm.

CONTACT DETAILS

You can reach Contributing Technical Editor David Marsh at forncett@btinternet.com.

REFERENCES

“Automotive fuse selection for transient overcurrent applications,” SAE technical paper no. 950299, www.sae.org. “Protection of automotive electrical systems,” section 3.8 of “General Technical Information: Applications”, Multilayer and SMD Varistors for Automotive Applications, www.epcos.com.