Power Electronics Handbook February 2017 by WTWH Media LLC

February 2017

Teardown: IKEAâ&#x20AC;&#x2122;s Smart LED bulb Page 8 Secrets of measuring currents above 50 A Page 34

Power Electronics HAND BOOK

Power Electronics 2-17 Cover V1.indd 1

2/24/17 11:32 AM

161101_ACCESS_EEW_US.indd 1 Digi-Key 12-16_Immediate Access.indd 1

11/1/16 2/16/17 3:35 4:24 PM PM

™

HIgH PERfORmaNCE BaTTERy HOLdERS

• For AA, AAA, 1/2AA and CR123A cylindrical batteries • Coil Spring and Polarized contacts assure low contact resistance and proper continuity for circuit protection • Available in Thru Hole Mount (THM) or Surface Mount (SMT) configurations • THM versions use nickel-plated stainless steel contacts • SMT versions feature gold-plated stainless steel contacts • Durable, rugged High Temperature Nylon housing • Quick & easy battery installation/replacement • Retains battery securely • Optional Covers are available for additional retention, if required • Ideal for low profile, space saving PCB applications • Suitable for Industrial and Consumer product applications • Request Catalog M65

It’s what’s on the InsIde that counts ® E L E C T R O N I C S

www.keyelco.com

EE-HB-THiNK Keystone PEH Corp_2-17.indd 2-17.indd 1 1

•

(718) 956-8900

C O R P.

•

(800) 221-5510

12/30/16 2/16/17 4:27 4:02 PM

POWER ELECTRONICS

HANDBOOK

The tough problems in power design

LEE TESCHLER EXECUTIVE EDITOR

AN ANNUAL RITUAL takes place in the upper echelons of the power electronics world. At the Applied Power Electronics Conference (APEC), considered a premier forum for applied power electronics, researchers present new ideas for power converters. Typically, these ideas are neither straightforward nor easy to understand. And there is generally nothing intuitive about them. So other researchers intrigued by these concepts spend a lot of time in the ensuing 12 months characterizing the most promising topologies and circuits mentioned at APEC. At APEC the following year, conference-goers can start to really understand how the original circuits behave when the new results are presented. This ritual is one of the reasons APEC has the reputation as a forum attended by practitioners who can analyze knotty technical problems. And in that regard, there are a couple of areas addressed in this year’s upcoming APEC that are worth mentioning. One of them concerns the promising field of siliconcarbide MOSFETs. Ljubisa Stevanovic, CTO of Silicon Carbide Works at GE Global Research, will talk about the SiC devices GE has devised for megawatt-scale industrial applications. “The toughest part of applying SiC is that it is not a one-for-one replacement for silicon IGBTs,” says Stevanovic. “So you can’t just remove the IGBTs from a solar inverter and plug in a SiC MOSFET. Instead of replacing an IGBT, you need to take advantage of the switching speed SiC transistors can provide.” One of things Stevanovic will cover at APEC is a solar inverter based on silicon IGBTs switching at 2.5 kHz that was recast with SiC MOSFETs operating at 8 kHz. The resulting inverter circuits were simpler

DESIGN WORLD — EE Network

commentary - power electronics hb 2-17 v1.indd 2

2 • 2017

partly because the SiC version only needed a twolevel bridge compared to the three-level design that the silicon version required. “IGBTs typically force you to use a three-level bridge. SiC devices can go up to ten times faster with about the same losses as IGBTs. That switching speed lets them operate with less filtering,” Stevanovic says. Attendees can also hear about SiC from Dr. Victor Veliadis, CTO and Interim Deputy Director of PowerAmerica, a DoE manufacturing institute working on advanced power conversion techniques. Veliadis and his PowerAmerica colleagues will be covering some of the difficulties they’ve managed to overcome in devising SiC devices and will discuss how SiC circuits excel compared to those using conventional silicon transistors. “A lot of the difficulty relates to getting processes that are specific to SiC and taking them from an R & D demonstration into a real manufacturing process,” he says. APEC’s reputation for discussions of tough converter topologies will also get some reinforcement from a presentation by Joel Steenis and Alex Dumais, two application engineers from Microchip Technology Inc. They will be covering how to devise a fully digital LLC resonant converter. “It is definitely a challenging topology,” says Dumais. “The biggest issue today is modeling it. Joel has done months of research on accurate modeling so engineers can design their own converters. The problem is the LLC circuit doesn’t behave in way that lends itself to modeling with state-space averaging, a quite common approach.” Dumais and Steenis say LLC converters are getting a lot of interest for powering servers and telecom systems because they are efficient and use fewer components than alternatives. These are just a few highlights. All indications are that APEC will not lack for discussions of knotty problems.

powerelectronictips.com | designworldonline.com

2/21/17 3:25 PM

MPD SPEAKERS

T. V

Gam ing console

P O R TA B L E V I D E O G A M E

D V D p l ay e r

cable box CELL PHONE

TA B L E T

GAME CONTROLLER

Help ing yo u en j oy the su per bowl this 2017

b at t e ry h o l d e r s . c o m Memory Protection Devices 1-17.indd 3

2/16/17 4:28 PM

POWER ELECTRONICS POWERELECTRONICTIPS.COM

EDITORIAL

DIGITAL MEDIA/MARKETING

VIDEO SERVICES

Editorial Director Paul J. Heney pheney@wtwhmedia.com @dw_Editor

Web Development Manager B. David Miyares dmiyares@wtwhmedia.com @wtwh_WebDave

Videographer Manager John Hansel jhansel@wtwhmedia.com @wtwh_Jhansel

Managing Editor Leslie Langnau llangnau@wtwhmedia.com @dw_3Dprinting

Digital Media Manager Patrick Curran pcurran@wtwhmedia.com @wtwhseopatrick

Videographer Bradley Voyten bvoyten@wtwhmedia.com @bv10wtwh

Executive Editor Leland Teschler lteschler@wtwhmedia.com @dw_LeeTeschler

Senior Web Developer Patrick Amigo pamigo@wtwhmedia.com @amigo_patrick

Videographer Derek Little dlittle@wtwhmedia.com

Senior Editor Miles Budimir mbudimir@wtwhmedia.com @dw_Motion

Web Production Associate Skylar Aubuchon saubuchon@wtwhmedia.com @skylar_aubuchon

FINANCE

Senior Editor Mary Gannon mgannon@wtwhmedia.com @dw_MaryGannon

Web Production & Reporting Associate Jennifer Calhoon jcalhoon@wtwhmedia.com @wtwh_Jennifer

Senior Editor Lisa Eitel leitel@wtwhmedia.com @dw_LisaEitel Associate Editor Mike Santora msantora@wtwhmedia.com @dw_MikeSantora

DESIGN & PRODUCTION SERVICES VP Creative Services Mark Rook mrook@wtwhmedia.com @wtwh_graphics Art Director Matthew Claney mclaney@wtwhmedia.com

@wtwh_designer Graphic Designer Allison Washko awashko@wtwhmedia.com

Traffic Manager Mary Heideloff mheideloff@wtwhmedia.com Production Associate Tracy Powers tpowers@wtwhmedia.com

Controller Brian Korsberg bkorsberg@wtwhmedia.com Accounts Receivable Specialist Jamila Milton jmilton@wtwhmedia.com

Digital Marketing Director Virginia Goulding vgoulding@wtwhmedia.com @wtwh_virginia Manager Webinars Stacy Combest scombest@wtwhmedia.com @wtwh_Stacy

2014 Winner

Marketing Manager, Social Media & Events Jen Kolasky jkolasky@wtwhmedia.com @wtwh_Jen 2011 - 2016

Marketing Coordinator Lexi Korsok lkorsok@wtwhmedia.com @wtwh_Lexi Digital Marketing Specialist Josh Breuler jbreuler@wtwhmedia.com @wtwh_Joshb Marketing Associate Aly Ryan aryan@wtwhmedia.com @wtwh_Aly

Director, Audience Development Bruce Sprague bsprague@wtwhmedia.com

find us on Twitter

@DESIGNWORLD

staff page -- power electronics 2-17 v1.indd 4

DESIGN WORLD — EE Network

2 • 2017

2/22/17 8:51 AM

A57_Layout 1 1/6/17 11:52 AM Page 1

The big name in miniature components. Transformers & Inductors, DC-DC Converters, AC-DC Power Supplies The most demanding applications require the world’s most reliable components. For almost 50 years PICO Electronics has been providing innovative COTS and custom solutions for Military, Commercial, Aerospace and Industrial applications. Our innovative miniature and sub-miniature components are unsurpassed in any industry. PICO Electronics’ products are proudly manufactured in the USA and are AS9100C Certified. To learn more about our products and how you can benefit from our expertise visit our website at picoelectronics.com or call us today at 800-431-1064.

• • • • •

Ultra Miniature Designs MIL-PRF 27/MIL-PRF 21038 DSCC Approved Manufacturing Audio/Pulse/Power/EMI Multiplex Models Available For Critical Applications, Pico Continues to Be the Industry Standard

.18"

ht.

TRANSFORMERS & INDUCTORS Think Pico Small Over 5000 std Ultra Miniature

think... low profile from

Surface Mount & Thru Hole Military • COTS • Industrial

DC-DC CONVERTERS

2V to 10,000 VDC Outputs — 1-300 Watt Modules • MIL/COTS/Industrial Models • Regulated/Isolated/Adjustable Programmable Standard Models • New High Input Voltages to 900VDC • AS9100C Facility/US Manufactured • Military Upgrades and Custom Modules Available

PICO Electronics Inc. 143 Sparks Ave, Pelham, NY 10803 Call Toll Free: 800-431-1064 E Mail: Info@picoelectronics.com • FAX: 914-738-8225

VISIT OUR EXCITING NEW WEBSITE: www.picoelectronics.com PICo PEH 2-17.indd 5

2/16/17 4:29 PM

inside

Cover photos courtesy of iStock.

THE POWER ELECTRONICS HANDBOOK

39 43

02 The tough problems in power design 08 Teardown: Inside Ikea's smart LED bulb

Press a button on the wall and a wireless connection tells this LED bulb what to do.

Modeling of losses in magnetic components Traditional ways of calculating the loss in inductors and other kinds of magnetic components have well-chronicled short comings that suppliers are beginning to address.

23 Better high-frequency power design through EM simulation

Here's how 3D EM simulation can be used to accurately gage gate-driver trace inductance.

28 The basics of thermistors in thermal management

Contents — PE Handbook 2-17 V1.indd 6

2 • 2017

Power measurements made with conventional meters can be drastically inaccurate. Here's how to decide when to bring in a specialized power analyzer.

34 Secrets of measuring currents above 50 A

There is an art to getting accurate readings of high current values. It pays to know a few of the tricks.

39 Power modules lessen the load on internal combustion engines

Integrated smart power modules will run mechanical loads that would otherwise sap fuel mileage.

43 How to be smart when selecting a fan for forced air cooling

Thermistors can reliably switch-in thermal protection to keep power electronics cool.

DESIGN WORLD — EE Network

30 The different routes to measuring electrical efficiency

It's easy to over-specify a cooling fan. But there are ways of keeping temperature down with fans that are sized just right.

powerelectronictips.com | designworldonline.com

2/22/17 10:37 AM

Sager EEWeb Design News 130 Anniversary.qxp_Layout 1 2/8/17 2:19 PM Page 1

It All Started in 1887 as a Small Boston Storefront Selling Speaking Tubes.

130 years later, Sager Electronics has grown into a leading North American distributor of interconnect, power and electromechanical electronic components. Acquired 5 years ago by TTI, Inc., a Berkshire Hathaway, Inc. company, Sager continues to operate with the same dedication to its core values – integrity, service and adaptability. Celebrating 130 Years of Distributing Confidence®

130 YEARS

1887 - 2017

Sager PEH 2-17.indd 7

2/16/17 4:30 PM

HANDBOOK

POWER ELECTRONICS

TEARDOWN:

Inside Ikea’s smart LED bulb LE LA N D TE S CHLE R | EX EC U T IV E EDIT OR

Press a button on the wall and a wireless connection tells this LED bulb what to do.

A NEW LED BULB from the furniture maker Ikea of Sweden is called the TRÅDFI. Now available in Europe, it’s expected to hit the U.S. soon. It comes with its own remote controller which mounts to the wall and uses a wireless connection to switch the bulb on and off, adjust bulb color temperature and dimming level. As smart LED bulbs go, this one is pretty simple. It doesn’t incorporate any sort of smart phone app and can’t be controlled over the Internet, like some other smart bulbs we’ve seen. The remote which controls the LED bulb has what seems like an elegant way of mounting to the wall. A base for the remote attaches to the wall either with screws or adhesive. Then the remote itself fits into the base, held on with a magnet. This scheme lets the user remove the remote to change its battery when the need arises and gives the wallmounted remote a sleek look. The battery is a CR 2032, a widely available 4.3-V lithium coin cell. The Ikea bulb talks to the remote in a way resembling that used by many other smart bulbs: The designers put the wireless chip on a separate circuit board that solders to the main board holding the driver electronics. In the case of the

The Ikea TRÅDFI LED bulb can be dimmed and colorcontrolled from a remote switch that communicates with the bulb wirelessly. The remote switch is magnetically attracted to a base (far right) that mounts on a wall. The magnetic attraction lets the switch be pulled down for replacement of its coin-cell battery.

DESIGN WORLD — EE Network

Ikea Teardown - PE Handbook 2.17 V2.indd 8

2 • 2017

powerelectronictips.com | designworldonline.com

2/21/17 3:47 PM

TEARDOWN: IKEA BULB

Ikea bulb, the wireless board mounts perpendicular to the main LED driver board with solder connections between the two on the edge of the wireless board. The identical wireless board also resides in the wall switch. In both the LED bulb and the wall remote, the wireless chip (EFR32MG1) is from Silicon Labs in Texas. This chip handles several wireless protocols that include BLE, ZigBee and Thread. Unfortunately we have no way to identify which of these the LED bulb uses. The chip is a system-on-a-chip device that includes a 32-bit ARM controller, a radio transceiver, I/O ports and serial interfaces, as well as numerous other peripheral functions. In fact, the chip integration is such that the only other items on the wireless board are a few passive components and a printed circuit antenna, which pokes out through a hole in the LED plate. Though the wireless board mounts on edge in the LED bulb, it sits flat against the main circuit board inside the wall-mounted bulb controller. The control’s main circuit board primarily holds the wireless board and the battery holder and serves as a substrate for the switch pad. The switch pad itself is a straightforward design where pressing a switch pushes down on a conductive rubber contact that, in turn, touches the circuit board to signal a button push.

Bulb architecture A point to note about the wireless board in the bulb is that there are only six soldered connections between it and the LED driver board. Assuming power and ground account for two of them, that means there are perhaps just four leads available to control the LEDs. In addition, there is no way to tell whether the wireless board on the remote is identical to that on the bulb; system-on-chip devices can be made with a variety of different options. The bulb has a physical layout that’s widely used for LED bulbs. It employs a 2.4-oz. heatsink that surrounds a circuit board potted in flexible epoxy-like material. The epoxy also provides some structural rigidity to the base of the bulb and the electrical contacts. The main circuit board and all the epoxy potting material fits inside a plastic sleeve that, in turn, slides inside the heat sink. The 28 LEDs on the bulb reside on a single metal plate that helps conduct heat and sits on the 2.4-oz. heat sink. Thermal grease sits at the interface between the LED plate and heat sink. Interestingly, this is the first LED bulb we’ve torn down that uses mechanical connectors, rather than just two soldered wires, for a connection from the circuit board to the LED plate. In fact, the Ikea

Not much to see. Visible here are the back of the remote control’s PCB containing the switches and the plastic wall switch containing the conductive rubber contacts. We removed one of the switch plates to show the conductive elastomer contact point.

powerelectronictips.com | designworldonline.com

Ikea Teardown - PE Handbook 2.17 V2.indd 9

2 • 2017

DESIGN WORLD — EE Network

2/16/17 7:47 AM

HANDBOOK

POWER ELECTRONICS

bulb uses dual connectors which each take the form of two pins extending from the main circuit board that mate with a simple contact on the LED board. We surmise the reason the bulb uses two connectors, rather than just one, is because the LEDs on the bulb are wired as two separate strings, one on each connector. That’s different than on most bulbs – most just wire all their LEDs in series in one long string. The rationale for two strings becomes clear from a close examination of the LEDs: There seem to be two different kinds. One style has a slightly darker amber color than the other. So the two strings evidently are wired so they can be operated independently. And that’s probably how the LED bulb manages dimming and changing its color temperature. It judiciously lowers or raises the current to either of the two strings to modulate both the amount of light output and probably the color temperature as well. However, the use of two LED strings presents a problem from the standpoint of driver design. Because most LED bulbs only use one string, most LED driver chips for use in LED bulbs can drive just one LED string. A bulb carrying two LED strings can’t easily use a driver chip designed to handle just one. Though there are numerous chips available for driving multiple strings of LEDs, the applications most of these chips address are in LCD backlighting or related display-type uses. They really aren’t set up for light bulbs. So they really aren’t appropriate for what Ikea is doing with this light bulb.

Silicon Labs EFR32MG1 Mighty Gecko Smart SoC

DESIGN WORLD — EE Network

Ikea Teardown - PE Handbook 2.17 V2.indd 10

2 • 2017

powerelectronictips.com | designworldonline.com

2/16/17 7:47 AM

TEARDOWN: IKEA BULB

Cutting off the acrylic bulb reveals the LED board. Visible here is the slot through which the wireless board antenna pokes through. Also evident are the two connectors used to mate the LED board with the bulb electronics – unusual in that most bulbs just use soldered connections. Also note LEDs that are light amber and dark amber. Dark and light LEDs are in separate strings, likely done as a way to implement color temperature changes and dimming. Removing the LED board (left) reveals the thermal grease used between the plate and the heat sink and the connector pins to the LED plate, coming from the epoxied electronics in the bulb base.

With the elastomer housing for the remote switch cut apart, we removed the PCB which mainly contains just the coin cell holder and Silicon Labs chip plus a few passive components.

powerelectronictips.com | designworldonline.com

Ikea Teardown - PE Handbook 2.17 V2.indd 11

2 • 2017

DESIGN WORLD — EE Network

2/16/17 7:51 AM

HANDBOOK

POWER ELECTRONICS

And that reality probably explains why the Ikea bulb doesn’t seem to contain a standard controller chip. There’s an unmarked eight-pin chip that seems to be handling LED control. Our guess is that it’s based on an LED driver design that was modified to handle two LED strings, but it’s source can’t be discerned from its markings. To analyze the LED drive, we can look at the other components on the circuit board and make educated guesses about the driver topology. Starting with the primary side of the circuit, we find five large capacitors near the diode bride and a power transistor, but no inductors. One possibility is that the transistor is there for purposes of power factor correction (PFC). And most LED bulbs these days do incorporate PFC, though we couldn’t find any power factor specs for the Ikea bulb. The fact that there are no big 60-Hz inductors in the primary side of the circuit tends to eliminate the possibility that PFC takes place passively. Most passive PFC circuits use bulky inductors, and it could be problematic to squeeze them into the base of an LED bulb along with everything else that must fit there. That’s why most LED bulbs go with active PFC if they do it at all. Moving past the circuitry near the base of the bulb, a large transformer resides on the circuit board that is likely part of the driver circuit itself. We might surmise from this setup that the transformer is there basically to isolate the secondary side of the circuit and to divide the input power into the two portions used by each of the LED strings. To determine the topology, we note that the circuit on the other side of the transformer consists of identical pairs of components as when there are two channels of LEDs. The fact that we find dual sets of these components tends to affirm the idea that the two LED strings are independently driven and controlled. The circuitry for each channel includes a power FET, an inductor, two sizeable capacitors, two beefy diodes, and a six-pin chip too small to be identified by its markings.

Ikea Teardown - PE Handbook 2.17 V2.indd 12

DESIGN WORLD — EE Network

2 • 2017

2/21/17 3:07 PM

Mean Well 1-17.indd 13

2/16/17 4:32 PM

POWER ELECTRONICS

HANDBOOK

Ikea Buck Circuits

Components on the bulb PCB might indicate that the LED driver topology is that of dual buck converters, one for each of the LED strings, driven through an isolation transformer.

Simple Passive PFC

The question becomes, what kind of a converter could employ these particular components? The most obvious answer would seem to be a buck converter. This topology steps down voltage while stepping up current. And it typically contains at least a diode and a transistor and at least one energy storage element such as an inductor. To reduce voltage ripple, it might incorporate a filter comprised of capacitors. So the buck converter would be our guess as to what we’re dealing with here. All in all, Ikea has put out an LED bulb that seems to work pretty well for those who only need something that responds to a nearby light switch. It’s not a candidate for those who want to dim or otherwise control a bulb over the cloud for some reason.

References Components found on the bulb PCB seem to indicate the bulb uses active rather than passive PFC.

DESIGN WORLD — EE Network

Ikea Teardown - PE Handbook 2.17 V2.indd 14

2 • 2017

Silicon Labs EFR32MG1 Mighty Gecko SoC www.silabs.com/Support%20Documents/TechnicalDocs/EFR32MG1-DataSheet.pdf

powerelectronictips.com | designworldonline.com

2/16/17 7:54 AM

TEARDOWN: IKEA BULB

Ikea board, top view

bottom view

The bulb PCB doesn’t contain a standard LED driver chip, but components found on the board seem to indicate the use of separate buck converters for each of the two LED strings, and for use of active power-factor correction. A side view of the main PCB (top) shows the wireless board. Its chip had metal shielding which has been removed in this image.

powerelectronictips.com | designworldonline.com

Ikea Teardown - PE Handbook 2.17 V2.indd 15

2 • 2017

DESIGN WORLD — EE Network

2/16/17 7:59 AM

POWER ELECTRONICS

HANDBOOK

Modeling of losses in magnetic components R A NJ I TH BR A M A N PA L L I | W U RT H EL EC T RON IK

Traditional ways of calculating the loss in inductors and other kinds of magnetic components have well-chronicled short comings that suppliers are beginning to address. IN A SWITCH MODE POWER SUPPLY (SMPS) most of the power losses take the form of switching and magnetic losses. Magnetic loss stems from the core and the windings in the storage/coupled Inductor. It has become more important to determine the inductor power loss accurately in the quest for higher energy efficiency and green technology. The estimation of core losses in a SMPS can require complex measurement set-ups, yet there’s no guarantee that the resulting estimates will apply in the particular application at hand. Historically, core losses are calculated using the Steinmetz equation and extensions of it. These equations can estimate losses reliably only for certain conditions or materials. Fortunately, new models have been developed to determine core losses effectively and accurately. In a SMPS, the Inductor acts as storage component. It stores energy in the form of a magnetic field during the switching-cycle ON time and delivers its energy to the load during the OFF time. Usually, an inductor consists of a coil predominantly made of copper wire and a core which has magnetic properties. In terms of electromagnetic physics, a magneto-motive force with respect to the time applied to the coil induces magnetic flux Ø(t). An important point is that at any location, the magnetic flux density B is always proportional to field intensity H.

A typical B-H curve when a sinusoidal excitation is applied to the core. The hysteresis loop area (red region) represents energy loss. Power loss depends on how many times per second the hysteresis loop is traversed. Thus, hysteresis loss varies directly with frequency.

where B is the magnetic flux density(Ø/A), μr is the permeability of the material, μo is the permeability of air, and H is the magnetic field Intensity. The coil is wound around or placed inside the core with an air gap to tap the magnetic field effectively. The core contains the air gap. The core is usually a ferrite material which has ferromagnetic properties and much higher permeability than air. Hence, the air gap places a high reluctance element of air in series with lowreluctance ferrite material, thereby locating the bulk of the energy in the air gap.

DESIGN WORLD — EE Network

Wurth - PE Handbook 2.17 V3.indd 16

2 • 2017

powerelectronictips.com | designworldonline.com

2/22/17 8:46 AM

MODELING OF LOSSES

Inductors operate according to the laws of Ampere and Faraday. Ampere’s Law relates current in the coil or turns of wire to the magnetic field in the core of the inductor. As an approximation, one assumes the magnetic field in the inductor’s core is uniform throughout the core length (le ). That assumption lets us write Ampere’s law as

where N is the number of turns of the coil around the inductor core and I is the inductor current. According to Faraday’s Law, voltage applied across the Inductor is

From above theories Inductor value is calculated as

Where Ac is core cross-sectional area. Because ferrite materials have high permeability, they offer an easy path for magnetic flux (low reluctance).

This characteristic helps contain the flux within the inductor’s core, which in turn enables the construction of inductors with high values and small size. This advantage is evident in the inductance equation above, in which a core material with a high permeability value allows for a smaller cross-sectional area. In a SMPS the peak magnetic flux density can be written as

Coupled inductors (sometimes flyback transformers) are also inductors but with multiple windings. The windings do have some special issues but their core properties remain same. Power dissipation in an Inductor happens in the windings and the core. These are losses and are termed winding losses and core losses. The power dissipation arises in windings due to the dc resistance (RDC) of the windings and because of phenomena such as skin effect and proximity effect. The loss due to dc resistance can be estimated using the equation:

Core loss graph for deriving constants

Core Loss graph plotted against peak flux density at different frequencies.

powerelectronictips.com | designworldonline.com

Wurth - PE Handbook 2.17 V3.indd 17

2 • 2017

DESIGN WORLD — EE Network

2/21/17 4:45 PM

The losses arising due to the skin and proximity effects can be termed the ac resistance (RAC) of the winding, which predominantly depends on operating frequency. There are a few techniques to determine these effects in magnetic components, but one would have to follow complicated procedures like Dowell’s method to estimate these losses. For core loss, B is measured as H is increased. The response of B versus H is nonlinear and exhibits hysteresis, hence the name hysteresis loop. Hysteresis is one of the core-material qualities that will cause power losses in the inductor core. Energy loss due to the changing magnetic energy in the core during a switching cycle equals the difference between magnetic energy put into the core during the on time and the magnetic energy extracted from the core during the off time. Using Ampere’s and Faraday’s Law, the energy in the core can be expressed as:

Energy lost in the core is the area traced out by the B-H loop multiplied by the core’s volume. The power loss is this energy (E) multiplied by the switching frequency. This expression applies as long as the core is not driven into saturation and the switching frequency lies in the intended (linear) operating range. The hysteresis loop area represents energy loss. Power loss depends on how many times per second the hysteresis loop is traversed. Thus, hysteresis loss varies directly with frequency. The second type of core loss is due to eddy currents, which are induced in the core material by a time-varying flux(dØ)/ dt) . According to Lenz’s Law, a changing flux induces a current that induces a flux in opposition to the initial flux. This eddy current flows in the conductive core material and produces losses of I 2 R.

Wurth - PE Handbook 2.17 V3.indd 18

DESIGN WORLD — EE Network

2 • 2017

2/21/17 3:11 PM

MODELING OF LOSSES

Determining losses Historically, core losses are estimated by using the power equation, also called Steinmetz equation.

To utilize the Steinmetz equation for nonsinusoidal waveforms in the estimation of core losses, a later extension, the Modified Steinmetz Equation (MSE) has been used widely.

(1) (2) where Pv is the core loss (hysteresis and eddy current loss) for unit volume, f is the frequency, Bpk is peak flux density of a sinusoidal excitation, K, α and β are the constants derived from core loss graph. The data given for core losses usually includes the effects of both hysteresis and core eddy currents. Core-loss measurements are difficult because they require complicated setups for measuring flux density and because they involve the estimation of hysteresis-loop areas. For plotting core loss charts, a sine wave is applied to a toroidal (ring) core with one or two one-turn coil windings. Then the constants are calculated using the graph of core loss graph plotted against peak flux density. The major drawback of the Steinmetz equation is that it is mostly valid for sinusoidal excitation. This is a huge drawback because in most power electronic applications, the inductor is usually exposed to non-sinusoidal flux waveforms. There are other models which separate hysteresis and eddy current loss to overcome the problem of non-sinusoidal waveforms. But the empirical Steinmetz equation has proven to be the most useful tool for sinusoidal flux-waveforms because it provides better accuracy and is quite simple to use. Hence there are extensions to this power equation to make it more accurate for non-sinusoidal flux waveforms.

feq being the equivalent frequency with respect to change in the duty cycle (DC) of nonsinusoidal waveforms. The MSE has some disadvantages. The Generalized Steinmetz Equation (GSE) was developed to overcome them:

As the GSE and MSE core loss charts are also based on sinusoidal excitation, they also have some limitations. And some core manufacturers also have devised their own models which work best with the cores they themselves manufacture. The major disadvantages of the Steinmetz and its extended models include the fact that they depend on empirical data of the core manufacturer (for core loss charts). This means the firm manufacturing the passive component

At left, typical modelling used for Steinmetz equation and its extensions. At right, Würth Elektronik’s point-of-operation approach.

powerelectronictips.com | designworldonline.com

Wurth - PE Handbook 2.17 V3.indd 19

2 • 2017

DESIGN WORLD — EE Network

2/22/17 9:23 AM

POWER ELECTRONICS

HANDBOOK

using the magnetic core has no control over the test setup, which can be quite complicated. These models also have a low accuracy with pulsating and triangular waveforms that characterize the SMPS. This is because the core loss graphs are created using the data from sinusoidal excitation. Similarly, extensions of Steinmetz models work best only for 50% duty cycle and limited frequency ranges. There are other problems. Due to the complexity of estimating magnetic path length, it is challenging to estimate core loss using existing models for iron powder materials and metal alloys, and the accuracy of these efforts varies widely. As well, Steinmetz models don’t factor in losses in the component winding caused by skin effects, proximity effects, and similar issues. Nor do these methodologies include ac losses of the windings. Finally, it is not practical to estimate losses in components which use more than one magnetic material. A new ac loss model Because of such difficulties, Würth Elektronik eiSos has developed a state-of-the-art model aimed at an easier optimization of systems containing inductors. The model is based on empirical data derived from real time application set ups. In the Würth model the total loss of the inductor is divided into two separate ac and dc losses. The power dissipation is due to dc current in the inductor windings and is listed as the dc loss.

The additional power loss due to ac flux swing in the coil and the core is termed the ac loss. To produce empirical data, a pulsating voltage is applied to a simple LC circuit and load resistance. The power input, Pin and power output Pout (across the capacitor in parallel with the load) are measured. Then PLoss=Pin-Pout is estimated and the ac loss of the Inductor PAC is separated out. This process is repeated over wide range of parameters including variation of peak flux density swing, frequency, ripple current, and so forth, to produce the empirical data. This empirical data is then used to create an equation to calculate ac loss in the form of PAC=f(∆I,freq, DC, k1, k2). The hysteresis loops shown in the typical core material data sheets represent the core overdriven by a sinusoidal waveform from + to – saturation. The hysteresis loop area represents energy loss. This is the same approach used to produce empirical data for the core loss charts. Of course, in an SMPS application, the core is usually driven by a much smaller rectangular waveform with peak flux density limited by core losses to a minor hysteresis loop. Power loss depends on how many times per second the hysteresis loop is traversed. Thus, hysteresis loss varies directly with frequency. The hysteresis loop changes shape somewhat with wave shape, current or voltage drive, and temperature. This variability, makes it hard to predict core loss accurately. The minor loop area depends on the voltage above the Inductor. Würth Elektronik’s ac loss model uses his minor loop in a point-of-operation approach to produce

REDEXPERT is Würth Elektronik’s online component selection and simulation tool for selecting power inductors. The user enters their input and output parameters in the desired topology, then the tool calculates the inductance value and displays the inductor choices. The Würth Elektronik ac loss model is implemented in this tool. With AC loss being calculated accurately, the estimation of temperature of a component is also apt with the application.

DESIGN WORLD — EE Network

Wurth - PE Handbook 2.17 V3.indd 20

2 • 2017

powerelectronictips.com | designworldonline.com

2/16/17 8:44 AM

MODELING OF LOSSES

Measured vs calculated ac losses

Pst, Pmse, and Pgse are core losses determined using the Steinmetz power equation, the Modified Steinmetz equation, and the Generalized Steinmetz equation respectively. Redexpert is the ac loss calculated using the Würth Elektronik ac loss model. Real is the measured ac loss.

powerelectronictips.com | designworldonline.com

Wurth - PE Handbook 2.17 V3.indd 21

2 • 2017

DESIGN WORLD — EE Network

2/21/17 4:46 PM

POWER ELECTRONICS

HANDBOOK

Typical dc/dc converter

empirical model data. This approach has proven to be robust and accurate over a wide range of parameters like frequency, ripple current and duty cycle. In Würth Elektronik’s ac loss model, empirical data is purely based on real-time parameters with accurate estimation of losses for any given duty cycle. The model is also accurate over a wide range of frequencies (10 kHz to 10 MHz) as the constants of the power equation are derived over a wide range with respect to the flux swing. And the model works for even small changes in core material and winding structure. Moreover, Würth Elektronik’s ac loss model is valid for components which have more than one material. It will accurately estimate losses of iron powder and metal alloy materials and is valid for any core shape and winding structure. And it includes dc winding losses. Würth Elektronik model has been validated and compared with existing models and measured data. The ac loss for various materials like WE-Super flux, iron powder, NiZn, MnZn, and so forth, are measured at a wide range of duty cycles, frequencies and other parameters compared with theoretical models. The Würth Elektronik’s ac loss model is accurate and practical. The model has been experimentally verified over a wide range of frequency, ripple currents and duty cycles and proved to be robust. This ac loss model is implemented in an online selection tool called Redexpert, eliminating the need for design engineers to request core loss charts.

An example illustrates how to use REDEXPERT. Consider a buck converter application. Suppose the user enters operational parameters of 8 – 12 V for the input voltage, 5 V for the output, and a switching frequency of 800 kHz, with inductor ripple at 40% and output current at 1 A. The REDEXPERT tool estimates the optimal inductance (Lopt) as 9.6 µH, Ton as 550 nsec and duty cycle as 0.44. It has also displayed more than 200 inductors as candidates. Assume the application requires a miniature and low-loss Inductor, and the user selects the WE-MAPI series. REDEXPERT has three topologies -- Buck, Boost and SEPIC – for which the user can select the component. Additionally, there is a loss calculator for a single inductor irrespective of the topology. REDEXPERT is web based so users needn’t bother downloading updates.

DESIGN WORLD — EE Network

Wurth - PE Handbook 2.17 V3.indd 22

2 • 2017

Setup of dc-dc converter for loss determination and resulting scope shots.

References Magnetics Design for Switching Power Supplies, by Lloyd H. Dixon. Available online from Texas Instruments, www.training.ti.com/ On the law of hysteresis, by C.P. Steinmetz. Available online from several sources including here: www.garagehacker.com/files/C.P%20 Steinmetz%20-%20On%20the%20 Law%20of%20Hysteresis%20-%20 Pt%201.pdf “Calculation of losses in ferro- and ferrimagnetic materials based on the modified Steinmetz equation” by Reinert, J.; Brockmeyer, A.; De Doncker, R.W., www.ieeexplore.ieee.org/ document/936396/ “Improved calculation of core loss with nonsinusoidal waveforms by Jieli Li; Abdallah, T.; Sullivan, C.R., www.wcmagnetics.com/wp-content/ uploads/2015/02/circulation.pdf REDEXPERT, www.we-online.com/redexpert

powerelectronictips.com | designworldonline.com

2/21/17 3:12 PM

BETTER HIGH-FREQUENCY POWER

Better high-frequency power design through EM simulation JOHN RIC E | TE X A S I N S TRU M E N TS I NC. .

Here’s how a 3D EM simulation can be used to accurately gage gate-driver trace inductance.

HIGH ELECTRON MOBILITY TRANSISTORS (HEMTS) were first introduced in the early 1980s and are touted for their exceptional switching qualities. These socalled heterojunction, field effect transistors (FETs) were originally developed for high-frequency RF power amplifiers. More recently, however, these heterojunction transistors have evolved to address a broader scope of high-frequency, high-voltage and high-current power applications. Today power transistors built on wide-bandgap (WBG) semiconductor materials including silicon carbide (SiC) and gallium nitride (GaN) are Gate charge compared being used to realize higher power efficiency and higher power density at a lower cost. In this regard, 3D electromagnetic (EM) field solvers help isolate and resolve printed circuit board (PCB) and component parasitic coupling that otherwise would compromise the wide bandgap (WBG) transistor advantage. MOSFETs built-in silicon and the latest WBG transistors are both FETs. As such, the turn-on and turn-off behavior is largely a function the device’s composite gate charge. Until the gate-drain “miller capacitance” of the FET is charged, the device remains in a high on-resistance state. The gate driver Gate charge of a 650-V enhancement mode GaN transistor sources and sinks current into the gate terminal to move up (green) and a super junction, silicon MOSFET (orange) of and down the device’s gate charge curve. The rate at which this comparable RDSON. Until the gate-drain “miller capacihappens is determined by the driver output current and the tance” of the FET is charged, the device remains in a high on-resistance state. device gate charge. It is the ultralow gate charge characteristic of WBG transistors that facilitates high-frequency operation in high-power applications, reducing both power stage and gate Voltage rise in a WBG buck converter driver losses in hard-switch applications. Also evident from a gate charge curve is that a full saturation of the vertically-structured, super-junction transistor requires nearly twice the gate potential compared with the laterally-structured GaN device. In fact, gate-source voltages exceeding about 6 V can damage some GaN transistors. GaN FETs with a low-voltage gate oxide capability need specialized GaN FET gate drivers like the Texas Instruments LM5113. As we will see, the key requirement in advancing highfrequency power is to minimize distributed parasitic coupling. To help address this need, TI introduced the LMG3410. This Switch-node voltage rise time (10-90%) transition to 48 V, 600-V, 12-A device integrates a gate driver with a depletion1.7 nsec for a 48 V to 1 V buck converter.

powerelectronictips.com | designworldonline.com

Texas Instruments - PE Handbook 2.17 V3.indd 23

2 • 2017

DESIGN WORLD — EE Network

2/21/17 3:13 PM

HANDBOOK

POWER ELECTRONICS

Typical distributed impedances

mode (normally on) GaN FET that delivers highfrequency, enhancementmode switch operation from a logic-level input. Some GaN manufacturers leverage the ultralow on-resistance of GaN technology by building high-voltage enhancementmode GaN transistors in a cascode arrangement of a low-voltage silicon MOSFET and highvoltage depletion-mode GaN device. These GaN devices typically exhibit a gate oxide capability comparable to that of silicon MOSFETs.

Distributed gate/driver impedances and sink/source currents for a half-bridge, gate-drive transmission line.

Current flow from the gate driver of a power FET is complex. It is affected by dc and ac parasitic impedances including PCB layout copper, component internal inductance, FET capacitances and driver dc resistance. If the gate driver output is overdamped, the GaN advantage What to expect with GaN will be partially lost. If the gate driver is underdamped, The accompanying figure illustrates the halfoscillations at the FET gate and switch node can create bridge switch node behavior of a 48-V-to-1-V excessive electromagnetic interference (EMI) and electrical buck converter delivering 10 A. At 24 V/nsec, overstress (EOS). the measured switch node dv/dt is at least ten The significant parasitic elements and gate drive current times faster than what is typically seen with silicon vectors are evident in the nearby figure depicting the of MOSFETs; and yet, GaN technology is capable of gate/driver distributed impedance and sink/source currents. even faster dv/dt. Still, the PCB and component But it is difficult to quantify the actual circuit behavior when parasitics that were benign with silicon MOSFETS parasitic elements are activated by switch-path di/dt and dv/ often become problematic when using WBG dt. Note that probing a switch node typically adds between transistors. 3 and 20 pF of capacitance and should be considered as an invasive measurement that can mask parasitic behavior. Measurement equipment Trace inductance as a function of vertical height including scope/probe bandwidth must be considered carefully. Now consider the parasitic elements of a transmission line. A transmission line is any pair of conductors used to move energy from one point to another. In the case of a MOSFET driver, the transmission line is from the driver decoupling capacitor, CVDD, to the FET gate terminal. It is essential to quantify the distributed impedance to ensure the resonant tank made up of the gate driver, PCB transmission line and the FET gate capacitance is properly damped. Trace Inductance of a 250-mil-long and Consider the inductance 50-mil-wide trace above a reference plane. of a trace as a function of the DESIGN WORLD — EE Network

Texas Instruments - PE Handbook 2.17 V3.indd 24

2 • 2017

powerelectronictips.com | designworldonline.com

2/16/17 9:37 AM

CONNECTIVITY SYNCHRONIZATION

P werShare Enabling Robust Automotive Solutions By Integrated Device Technology

SAFETY • Seat Occupant Detection • Hands-Off Detection

INFOTAINMENT • Wireless Charging • Timing and Signal Conditioning • Advanced Memory • Power Management • Environmental Sensors

BODY • HVAC • Interior Lighting • Power Seat • RainASensor smartphone you can charge without wires is not news. • Sun Sensor But using that same smartphone to wirelessly charge your smartwatch? • Humidity Sensor • Mirror Control

POWERTRAIN • Transmission • Throttle Control • Manifold Pressure/ Temperature • Oil (Level, Temperature, Pressure, Quality) • In-Cylinder Pressure • UREA Pressure ™ • Pressure Sensor • Mass Air Flow • E-Gas • Diesel Heating • HTS • Flex Fuel

That’s the power of IDT’s Wireless PowerShare . IDT – The Leader in Wireless Power Technology IDT.com/go/powershare

IDT PEH 2-17.indd 25

2/16/17 4:31 PM

HANDBOOK

POWER ELECTRONICS

vertical height, h, above the reference plane. At h = 30 mil above the reference plane, the trace would have about 750 pH of inductance. Reducing the height, h, to 3 mils would result in a proportional reduction of one-tenth the inductance or 75 pH. Clearly, minimizing stack-up height is essential for reducing PCB trace inductance. Note this first-order approximation does not account for the influence of other material including co-planar copper, copper weight and material dielectric. It takes a finite element analysis (FEA) to precisely analyze the frequency-dependent PCB trace impedance. To solve for the broadband surfacecurrent interaction requires the application of Maxwell’s equations to the finite element mesh. Fortunately, software and computing power has advanced to the point where this kind of analysis is now possible, even on a laptop computer. More sophisticated analysis will become increasingly important in high-frequency power applications to realize optimal and reliable circuit behavior. EM simulation tools like CST Microwave Studio can quantify the frequency dependent, parasitic loop inductance in a gate drive circuit. Consider the following example analysis conducted on an Efficient Power Conversion (EPC) EPC920X evaluation module. The EPC920X half-bridge power stage layout and schematic were both optimized by EPC engineers who are well aware of GaN transistor

At left, plots of the entire loop inductance with and without R24, a 0402, zero-Ohm gate resistor. At right, the partial inductance of the segment illustrated in the board mesh of low-side gate sink drive. It is important to point out that the physical dimensions of any discrete components (R24 in this case) and even the analysis ports will affect the inductance calculation.

DESIGN WORLD — EE Network

Texas Instruments - PE Handbook 2.17 V3.indd 26

2 • 2017

EPC EPC920X half-bridge power stage PCB layout (a); and schematic (b).

PCB import into CST Microwave Studio (a); small section associated with low-side drive (b).

Board mesh of low-side gate sink drive (a); and port assignment to calculate partial inductance (b) for the EPC920X half-bridge power stage.

powerelectronictips.com | designworldonline.com

2/21/17 4:00 PM

BETTER HIGH-FREQUENCY POWER

An illustration of how “ports” are assigned to all relevant copper and components, including the LM5113 gate driver, GaN transistors and discrete components. With this assignment complete, SPICE/IBIS models (or scope captured waveforms) can then be used to excite and interact with the EM-simulated PCB to determine high-frequency circuit behavior. In this way all distributed PCB and component parasitic elements are considered; you can even determine the PCB electromagnetic interference (EMI) patterns

switching speeds; many designers who want to use GaN technology won’t have the same kind of sensitivities. A 3D EM simulation of a PCB section of the PCB helps to precisely calculate the gate drive loop inductance. The analysis here focuses on the interconnect impedance associated with the low-side copper from the gate of Q2 through R24 to the LM5113 gate driver sink pin. The board design was exported from the Altium Designer layout tool as an OBD++ file and imported into CST Microwave Studio. OBD++ files export all physical properties of the board including the layer stack-up. To minimize analysis run-time, we mesh only the relevant portion of a PCB, in this case the lowside gate driver. The 3D EM simulation of the section took four minutes to run on a laptop computer with four cores. To calculate partial inductance, we assign current ports to various copper segments of the loop. After meshing and solving the EM coupling mechanisms, we apply a broadband current excitation from 500 kHz to 500 MHz to the gate drive loop to determine the inductance. The real power of a 3D EM simulator comes into play with an EM analysis of the entire PCB. The board can then be excited within the EM simulator environment using waveforms representative of the application. This “co-simulation” links all the physical coupling mechanisms of the PCB with a circuit-level SPICE analysis. In this way an accurate assessment the board’s high-frequency circuit behavior can be realized. The learning curve of any simulation tool can be steep; this is especially true with 3D EM simulation. Application of these tools requires a deep understanding of the EM simulator and the end application. If possible, it’s best to pick a tool that has an intuitive user interface, practical application material and excellent technical support. I found all of these at Computer Simulation Technology (CST) and especially appreciated the excellent technical support I received from Patrick DeRoy.

powerelectronictips.com | designworldonline.com

Texas Instruments - PE Handbook 2.17 V3.indd 27

References LMG3410 data sheet http://www.ti.com/lit/ds/symlink/lmg3410.pdf Rice, John. Are you accurately measuring the picosecond rise time of your GaN device?, Power House TI E2E Community Blog, July 1, 2015 www.e2e.ti.com/blogs_/b/powerhouse/ archive/2015/07/01/are-you-accuratelymeasuring-the-picosecond-rise-time-of-yourgan-device How to Drive GaN Enhancement Mode Power Switching Transistors, GaN Systems Application note, GN001 Rev 2014-10-21 www.tinyurl.com/js2ymon CST Microwave Studio www.cst.com/Products/CSTMWS Rice, John. Get into electromagnetic compliance with GaN. TI E2E Community Blog, August 26, 2015 www.e2e.ti.com/blogs_/b/powerhouse/ archive/2015/08/26/get-into-electromagneticcompliance-with-gan Getting Ahead with Coupled 3D Field and Circuit Simulation, Computer Simulation Technology (CST) Webinar www.cst.com/events/webinars/2016-02-18field-circuit-simulation

2 • 2017

DESIGN WORLD — EE Network

2/16/17 9:37 AM

POWER ELECTRONICS

HANDBOOK

The basics of thermistors in thermal management S ONJA BROWN | E P C OS , A T DK G ROU P C O.

Thermistors can reliably switch-in thermal protection to keep power electronics cool. ALL SEMICONDUCTORS generate thermal losses, of course. Thermal losses in power semiconductors can range from a few watts to multiple kilowatts. In heatsinks, the conductive capability is typically measured in °K/W. The smaller the conductive capability value, the better the thermal dissipation capacity. Three entities are necessary to determine the right heatsink for a power semiconductor: the thermal contact resistance, the maximum heat dissipation in the semiconductor, and the highest expected ambient temperature.

Some thermal dissipation happens by convection, but the limits of this method are quickly reached – especially with small chips that dissipate a lot of power. Consequently, it is impossible to guarantee – and not wise to rely on – convection cooling. Furthermore, heatsinks are often too big for compact designs. The best course of action is to actively dissipate heat via fans or combined air and water-cooled heat exchangers, which are often operated without regulation. In many electronic power supplies and converters, power dissipation is loadrelated. To improve the energy balance and minimize the generation of noise, the active thermal dissipation switches on only when the electronics hits a defined temperature limit.

Typical PTC resistance behavior

Thermistors are good candidates for detecting temperature limits. Thermistors can be positive temperature coefficient (PTC) or negative temperature coefficient (NTC) devices. With NTC thermistors, resistance falls with rising temperature. In PTC thermistors, the resistance increases as temperature drops. On exceeding a specific temperature, PTC thermistors show a sharp rise in resistance, making them useful as temperature

Typical NTC resistance behavior

PTC thermistors have a sharp rise in resistance with temperature. NTC thermistors have more linearity.

DESIGN WORLD — EE Network

TDK.EPCOS - PE Handbook 2.17 V1.indd 28

2 • 2017

powerelectronictips.com | designworldonline.com

2/21/17 3:14 PM

THE BASICS OF THERMISTORS

Example: Audio circuit with temperature protection A complete high-performance audio end-stage circuit with double temperature protection. This circuit can monitor the heat of four heatsinks. Above 85°C the fan activates. If temperature ever reaches 100°C, the dc voltage protection circuit receives a positive signal, causing its relay to trip and disconnect the load.

limit sensors. NTC thermistors, on the other hand, exhibit greater linearity and are therefore suitable for temperature measurement. Because of their steep temperature-resistance curves, PTC thermistors can monitor temperature limits and actuate a fan when temperature is high enough. PTC thermistors can connect in series and when functioning as temperature sensors can monitor several hot spots. When a PTC sensor in a series connection exceeds the specified limit temperature, it switches the circuit to the high-ohmic state. As an example, this action can be used in notebooks to monitor the main processor, the graphics processor and any other heat-emitting components using SMD PTC sensors. Another application of PTC sensors is in the thermal monitoring of motor windings. For example, some manufacturers offer special types of PTC sensors that can be easily integrated into the windings of three-phase motors. NTC thermistors can also be used for temperature monitoring – particularly when the linearity of their temperature-resistance characteristic comes in handy. For example, a high-performance audio end stage might use NTC sensors to detect two temperatures. In such a design, eight output transistors with emitter resistors can mount together on a cooling fan unit to minimize the housing dimensions. Four separate heatsinks might be arranged point-symmetrically to one another, with two power transistors on each heatsink. Because the output transistors sit on four heatsinks that are both electrically and thermally insulated from one another, each heatsink must be monitored individually. This individual monitoring is necessary because even if the heat sinks are adequately dimensioned, the tolerances can result in a slightly uneven load distribution. The thermal monitoring must let the circuit take two actions: First, as soon as one or more heatsinks reaches a temperature of 85°C, the fan must switch on. Second, a load must be shed on reaching a temperature of about 100°C. NTC sensors available from certain manufacturers are useful in scenarios where space is a constraint. They let a single NTC temperature sensor do the job rather than two sensors. The design outlined above lets the fan switch on reliably at 85 °C. As this design is relatively slow in thermal terms, there is no need to build in an on-off hysteresis circuit as would normally be the case. A wide range of NTC and PTC sensors are available. These sensors can perform effectively and reliably in thermal management systems for power electronics.

powerelectronictips.com | designworldonline.com

TDK.EPCOS - PE Handbook 2.17 V1.indd 29

References EPCOS, a TDK Group Co. www.epcos.com

Examples of EPCOS PTC sensors include (from left to right) an SMT-PTC temperature sensor for mounting on printed circuit boards; PTC sensors for integration in motor windings; and a PTC sensor with lug for mounting onto heatsinks.

2 • 2017

DESIGN WORLD — EE Network

2/21/17 4:02 PM

POWER ELECTRONICS

HANDBOOK

The different routes to measuring electrical efficiency B OB Z OL L O | K EYS IG H T T E C H NOLO GI ES, I NC.

Power measurements made with conventional meters can be drastically inaccurate. Here’s how to decide when to bring in a specialized power analyzer. IT’S IMPORTANT to measure energy efficiency accurately in applications running from limited power sources such as batteries. And an understanding of how a device consumes energy can lead to a better understanding of operational costs regardless of whether the source of energy is limited or not. Server farms, for example, consume enough energy to influence power consumption on a national scale, and the operational costs of the server farm are set to a great degree by the cost of the power to run the servers. High accuracy is often a requirement in measurements of energy efficiency. For example, developers may have to detect small changes in power consumption to determine if a new design hits overall device efficiency goals. As efficiency approaches 90% to 100%, the measurement of

overall efficiency becomes quite challenging as it becomes necessary to measure tiny differences in power-in and power-out to determine the energy lost. First a few basics. Electrical power is usually calculated by measuring both voltage and current and then calculating the product of the two. If the power is pure dc, this calculation is fairly straightforward. The measurement becomes more complex with ac power because it involves phase relationships, and often engineers turn to power analyzers for these measurements. But regardless of the

Power Analyzers such as the Keysight PA2200 Series IntegraVision power analyzers easily measure power conversion efficiency. They also provide dynamic measurement capability to capture and display transient voltage, current, and power waveforms.

DESIGN WORLD — EE Network

Keysight - PE Handbook 2.17 V3.indd 30

2 • 2017

powerelectronictips.com | designworldonline.com

2/16/17 10:22 AM

THE DIFFERENT ROUTES

measurement tool (DMM, voltmeter, ammeter, scope, power analyzer), power measurements must take place under the right conditions. Suppose we measure the power a light bulb consumes. We can measure the ac power in, and this entity will be static (assuming we are not measuring an electronically dimmable LED bulb). We can first measure RMS voltage on the bulb, then the RMS current into the bulb, then multiply the two entities together and arrive at the right power consumption, as the RMS voltage and RMS current consumption won’t change vs time. This is not the case for the power consumed by a washing machine. The washer cycles between pumping, soaking, agitating, and spinning. Each stage draws a different amount of power. So the manufacturer will probably average or integrate the power consumed over a full wash cycle to come up with an average consumption figure. This practice is probably OK as this cycle is predictable and repeatable. But other devices, like electric vehicles or cell phones, consume power in ways that are unpredictable and based on the behavior of the individual user. In the case of these dynamic power consumers, the usual approach to measuring power consumption is to decide on a specific power consumption profile and only then settle on a measurement method. The reason is the power profile of devices in the real world can be complicated enough to demand specialized measurement methods. Suppose, for example, we use a DMM to measure RMS voltage and then switch over to ammeter mode and measure RMS current. The current will be measured at a different time than the voltage. We won’t get an accurate calculation of power if the voltage and current are changing. Even relatively simple devices increasingly require a simultaneous voltage and current measurement to calculate instantaneous power accurately. 2 • 2017

DESIGN WORLD — EE Network

Keysight - PE Handbook 2.17 V3.indd 31

2/16/17 10:24 AM

POWER ELECTRONICS

HANDBOOK

The instantaneous measurements are then integrated or averaged over the time interval of interest to determine the overall power consumption. So much for dynamics on a macro level, where it is obvious the power is varying. Now consider power on a microscale. Most power converters today are switch-mode devices (or switching power supplies). The input voltage to the switch-mode converter can be constant (dc or ac RMS), but the current will not be. The switch-mode power supply draws current nonlinearly – it does not look like a resistive load. It draws current in pulses if running from dc or in the form of non-sinusoid waveforms if running from ac. These nonsinusoidal waveforms make it difficult to calculate the instantaneous power accurately. To properly measure power, you must simultaneously capture voltage and current, point by point, multiply them together to get an instantaneous power, then integrate the resulting relationship to determine the real power consumption. Effectively, a power analyzer goes through the above steps. That is why a power analyzer is the weapon of choice when measuring power. Regardless of the wave

shape or skew of the voltage or current waveforms, the answer will be correct as long as you are within the bandwidth of the power analyzer (typically 10 kHz or higher).

Measuring static dc/dc converter efficiency Under static conditions (basically, a constant load and supply of power), it is relatively easy to measure the efficiency of a dc/dc converter. The inputs and outputs are dc, and if they not changing, then a DMM as a voltmeter can measure the input voltage. You can also use the DMM as an ammeter to measure average input current and calculate average input power. Similarly, a DMM can measure output voltage and current. With all measurements in steady state and in the form of dc, there is no concern about time skew between the voltage and the current measurement or between input and output power measurement. DMMs can offer outstanding measurement accuracy so the accuracy of efficiency measurements based on their readings will be high as well. For example, in a Keysight 34460A DMM, the basic measurement accuracy of 0.0025% for dc voltage and 0.007% for dc current. The efficiency calculations based on these readings can be on the order of 0.1%, depending on the voltage and current range. Now consider the case where conditions on the dc/dc converter change, as with a variable input voltage or variable load current. Now you must treat the converter as an ac device. When inputs or outputs are time varying, the DMM method will not always work. Voltage and current will most likely not be in phase. An RMS voltmeter can probably measure ac input voltage accurately, as it is most likely a sine wave. The current waveform will most likely not be

A screenshot from a Keysight IntegraVision power analyzer shows the waveforms into and out of a 225-W dc-to-ac solar inverter. On the right, the analyzer is making a power efficiency measurement. Channel 1 is measuring the dc in from the solar panel (30 V, 7.57 A, 235 W). Note the dc is not pure dc and has some ac ripple, which is properly accounted for in the Channel 1 power measurement done by the power analyzer. Channel 2 is measuring the ac out to the grid (242 Vac rms, 933 mAac rms, 225 W). The resultant calculated efficiency is 96.05%.

DESIGN WORLD — EE Network

Keysight - PE Handbook 2.17 V3.indd 32

2 • 2017

powerelectronictips.com | designworldonline.com

2/16/17 10:29 AM

THE DIFFERENT ROUTES

a sine wave. It may have a high crest factor or other non-sinusoidal qualities, thus making RMS current measurements challenging. A four-channel oscilloscope can be a good tool for measuring efficiency. Most certainly, all four channels will simultaneously measure input voltage, input current, output voltage, and output current. However, oscilloscopes are typically ground referenced, which could pose a challenge for measuring some power converters. But the problem can be addressed through use of differential voltage probes and current probes. However, the oscilloscope does have some limitations. In general, an oscilloscope has about 3% error in its vertical measurements due to both the oscilloscope gain error and the external probe error. This 3% error will cause the oscilloscope to struggle to accurately measure efficiencies over 90% or to see incremental changes of 5% or less in efficiency. For measuring lower-efficiency devices, or to get a general sense for the efficiency, an oscilloscope is a handy tool.

Use of a power analyzer is a preferred method to measure power converter efficiency. These instruments are designed specifically to make accurate, simultaneous measurements of Keysight PA2200 Series IntegraVision Power Analyzer, voltage and current. Multi-channel www.literature.cdn.keysight.com/litweb/ analyzers can simultaneously pdf/5992-0324EN.pdf?id=2544198 measure both input and output power on a single-phase or multiphase ac or dc signal. Most power analyzers are floating, capable of measuring hundreds of volts without differential probes, and can measure current directly. Power analyzers such as the Keysight PA2200 Series IntegraVision Power Analyzer easily measure power conversion efficiency with high accuracy. With a power analyzer, you can expect to measure power conversion efficiency with 0.25% to 0.5% accuracy. Power analyzers also handle the complexity of multiphase ac power measurement.

References

Industry's Lowest On-Resistance Ultra-Junction MOSFETs at 650V and 850V Enabling Very High Power Density

Part Number IXFB150N65X2 IXFN150N65X2 IXFN170N65X2 IXFB90N85X IXFN90N85X IXFN110N85X

Features: • • • • •

VDSS (V) 650 650 650 850 850 850

RDS(on) max. TJ=25°C (mΩ) 17 17 13 41 41 33

Ultra low on-resistance RDS(on) and gate charge Qg Fast body diode Superior dv/dt ruggedness Avalanche capability Low package inductance

Qg typ.

EAS

dv/dt

(nC) 355 355 434 340 340 425

(J) 4 4 5 4 4 3

(V/ns) 50 50 50 50 50 50

Package Type PLUS264™ SOT-227 SOT-227 PLUS264™ SOT-227 SOT-227

SOT-227

Applications: • • • • • • •

High-eﬃciency switched-mode and resonant-mode power supplies Electric vehicle battery chargers AC and DC motor drives DC-DC converters Robotics and servo control Power Factor Correction (PFC) circuits Renewable energy inverters EUROPE IXYS GmbH marcom@ixys.de +49 (0) 6206-503-249

USA IXYS Power sales@ixys.com +1 408-457-9042

PLUS264

ASIA IXYS Taiwan/IXYS Korea sales@ixys.com.tw sales@ixyskorea.com

www.ixys.com

Keysight - PE Handbook 2.17 V3.indd 33

2/16/17 10:30 AM

HANDBOOK

POWER ELECTRONICS

Secrets of measuring currents above 50 A G E ORG E S E L BACHA , EVA N S HORM A N, H A RRY C H A N DRA | A L L EG RO M IC ROS YS T E M S L L C

There is an art to getting accurate readings of high current values. It pays to know a few of the tricks.

IT CAN BE CHALLENGING to sense currents exceeding 50 A because the task often involves thermal management, must take place in a limited PCB area, and in some cases, requires a voltage isolation device. Two widely used methods for sensing high currents are a sense-resistor/op-amp approach, and Hall-based current sensing. It is useful to compare these two techniques. Recently developed Allegro MicroSystems integrated current sensors, ACS780LR and ACS770CB, will be used as examples. Traditionally, the typical way to measure current is by putting a sense resistor in series with the current-carrying conductor. An op amp may then measure the voltage drop across the resistor, and the current is calculated from Ohm’s law. The nominal resistance of the sense resistor is selected to supply a sufficiently large voltage while also minimizing the power dissipation (P = I2R). Large currents will require a low resistor value (usually 1 to 50 mΩ) to minimize the heat generated on the board. However, too-low resistor values will lead to a small sensed voltage and, in turn, a low-accuracy measurement. Low resistor values will also have a large footprint that consumes valuable PCB area. The thermal coefficient of the sense resistor and the voltage offset of the op-amp will also contribute to the measurement error. As a result, designers must balance among accuracy, power consumption, thermal management, PCB area, and cost. It’s often best to measure current near the supply voltage of the load (the high-side) instead of near ground (the low-side). Measuring on the high-side brings immunity to ground bounces and allows for the detection of short circuits to ground. Depending on the supply voltage and the application, basic or reinforced isolation might be needed for sense-circuit connections. If a sense-resistor/op-amp are used to measure on the highside, an op-amp with a high common-mode input range will be necessary, making the design more complex. To provide isolation, additional isolators (such as optocouplers) and isolated power supplies will be needed, increasing complexity and boosting costs. 34

DESIGN WORLD — EE Network

Allegro - PE Handbook 2.17 V3 - update for digital edition.indd 34

2 • 2017

The ACS770 in the CB Package (left) and ACS780 in the LR package (right). A phantom view of the ACS780 in the LR Package shows how as the current flows into the integrated conductor, a magnetic field is generated and sensed by the on-die Hall elements. In the CB package, current flowing in the integrated conductor generates a magnetic field that is concentrated by the integrated low-hysteresis core and sensed by the Hall element. The core also acts as a magnetic shield, rejecting external stray fields

powerelectronictips.com | designworldonline.com

3/3/17 2:20 PM

SECRETS OF MEASURING CURRENTS

On the other hand, Hall-effect current sensor ICs, such as those provided by Allegro, eliminate the need for a sense-resistor. The current flows directly into the integrated conductor, generating a magnetic field that will be measured. The ACS780 sits in a 6.4 × 6.4-mm surface mount LR package. Current flows into the integrated conductor and generates a magnetic field that on-die Hall elements then sense. Use of a flip-chip assembly technique brings the Hall elements quite close to the leadframe where the magnetic field is at its highest point. This packaging allows for superior signal-to-noise ratio. The device uses two Hall elements to detect and reject any external stray magnetic field. The integrated conductor has a low 200-µΩ resistance to reduce power dissipation, allowing for more than 100 A of continuous current measurement with a 120-kHz bandwidth. Thermal performance depends highly on PCB design and layout The ACS770 sits in a 14 × 21.9-mm through-hole CB package. As current flows in its integrated conductor, an integrated lowhysteresis core concentrates the magnetic field which is then sensed by the Hall element with a typical accuracy of ±1% and 120 kHz bandwidth. The core also acts as a magnetic shield, rejecting external stray fields. The integrated conductor has 100-μΩ resistance, providing ultralow power loss. The ACS770 can measure 200 A continuously at an ambient of 85°C and can be factory programmed to measure inrush currents up to 400 A.

Via count: 100% (original design)

Via count: 50%

Thermal via patterns of the ACS780LR evaluation board. Reducing the number of thermal vias by 50% (18 instead of 36) brought a 5.6°C rise to 156°C. Removing all thermal vias brought a 33.5°C rise to 183.5°C.

CB OVERTEMP WITH APPLIED DC CURRENT

Thermal performance To determine the appropriate sensor for an application, it’s important to understand the thermal performance under high-level transient currents and constant dc/RMS currents. For the examples that follow, all measurements took place at 25°C ambient and could be used to derate the sensors at different operating temperatures. High-current pulse testing of the LR package took place using the Allegro ACS780 evaluation board. This is an eight-layer board powerelectronictips.com | designworldonline.com

Allegro - PE Handbook 2.17 V3 - update for digital edition.indd 35

2 • 2017

DESIGN WORLD — EE Network

3/3/17 2:20 PM

HANDBOOK

POWER ELECTRONICS

Temperature rise in current-sensors with two-ounce (70 μm) copper and an FR4 substrate. Thirty-six thermal vias of 0.2 mm diameter were placed next to each of the solder pads of the integrated current conductor. The package then experienced a current pulse of a set magnitude and the time was measured for two conditions: the time for the die temperature to exceed the maximum junction temperature of 165°C, and the time to fuse the current conductor open. All testing of the CB package took place using the Allegro ACS770 evaluation board. This is a two-layer board with four-ounce (140 μm) copper and an FR4 substrate. Sixteen thermal vias of 0.5 mm diameter were placed next to each of the solder pads of the integrated current conductor. When subjected to high-current pulse testing, the CB package did not fuse at 1.2 kA—the maximum current capability of the lab equipment performing this measurement. The nearby table shows the maximum current pulse duration and duty cycle that could be applied to remain within the safe operating zone where the die temperature of 165°C is not exceeded. A nearby figure shows die temperature rises as continuous dc current is injected through the sensors and temperature reaches steady state. As expected, the CB package shows a smaller temperature increase because of its lower conductor resistance of 100 μΩ compared to 200 μΩ for the LR package. The system thermal performance depends greatly on the PCB layout and can be improved in several ways: by incorporating multiple layers of metal to better dissipate the heat under the IC, by adding a heat sink as close as possible to the IC, or

Here’s how the die temperature rises in the two packages as continuous dc is injected through the sensors and temperature reaches steady state.

Safe operation zone

LR package fuse and overtemperature time as a function of applied dc current. The green area shows the safe operating zone where the die temperature remains below 165°C. The orange area shows the conditions under which the maximum junction temperature is exceeded but the current conductor is not fused.

A view of the ACS780 evaluation board. The thermal via pattern used for the solder pads of the integrated current conductor is at left.

DESIGN WORLD — EE Network

Allegro - PE Handbook 2.17 V3 - update for digital edition.indd 36

2 • 2017

powerelectronictips.com | designworldonline.com

3/3/17 2:21 PM

SECRETS OF MEASURING CURRENTS

The Allegro ACS770 evaluation board. The thermal via pattern for the solder pads of the integrated current conductor is at right.

by adding thermal vias (that connect all metal layers) surrounding the Allegro IC integrated conductor solder pads. Of the three methods, adding thermal vias has the least impact on PCB area and cost, and it is easy to implement. To understand the impact of vias and how many to use, a simulation was run on the ACS780LR evaluation board using a natural convection model. The model assumed an air enclosure of 300×300×300 mm with the enclosure wall set to 25°C. The injection of current caused the steady-state die temperature to reach 150°C. Reducing the number of thermal vias by 50% (18 instead of 36 vias per solder pad), brought a 5.6°C rise in die temperature to 156°C. Removing all thermal vias caused a 33.5°C rise in die temperature to 183.5°C. These results highlight the significant benefits of having thermal vias, while showing that a small reduction in the number of vias (much less than a 50% reduction relative to the Allegro evaluation board) should have a minimal impact on the thermal

performance. The small footprint of the ACS780 in the LR package and its ease of surfacemount assembly brings advantages for measuring currents exceeding 100 A. The approach is to reroute a portion of the current to be sensed through a trace on the PCB. Thus, a portion of the current to be sensed does not pass through the Allegro IC. Here, the current ratio of the splitter is critical. It must be set so the maximum possible current flows through the sensor (while the sensor remains in the thermal safe operating zone) to get the best accuracy. A simulation illustrates the thermal capability of this approach. Suppose we have a board with a current ratio of 6.7:1 (that is, current through trace: current through sensor) and the following specifications: six copper layers (top and bottom layer thickness of two-ounce (70 μm), inner layers of three-ounce (105 μm)), an FR4 substrate, 36 thermal vias of

An illustration of the current splitting principle using the ACS780LR. At right is the layout of the current splitting board used in the thermal simulation of the ACS780LR package.

powerelectronictips.com | designworldonline.com

Allegro - PE Handbook 2.17 V3 - update for digital edition.indd 37

2 • 2017

DESIGN WORLD — EE Network

3/3/17 2:21 PM

HANDBOOK

POWER ELECTRONICS

Comparing op amp sensing and discrete current sensors

0.2-mm diameter around each pad, and 5-mm diameter through-holes for current injection on the PCB. An aluminum heat spreader of 94 × 70 mm connects under the PCB. With 250 A injected in the PCB, a simulation assumed natural convection with an air enclosure volume of 300×300×300 mm with the enclosure wall set to 25°C. The highest observed temperature was 74°C on the top metal (~50°C rise relative to the ambient temperature), while the die temperature reached 71°C. Allegro current sensors are galvanically isolated, offering an efficient way to measure on the high-side. The ACS780LR targets applications where the supply voltage is less than 100 V. Its construction provides inherent isolation, because the active circuitry on the die is not electrically connected to the current conductor. The ACS770 is certified to UL 60950-1 2nd edition, passing 4.8 kV for 60 sec. Its basic Isolation working voltage is 990 (Vpk or dc) or 700 Vrms, while its reinforced isolation working voltage is 636 (Vpk or dc) or 450 Vrms. All in all, advances in packaging and circuit design have simplified the task of using Hall current sensor ICs to measure currents exceeding 50 A on a PCB. Accurate and galvanically isolated sensing can take place economically with little power loss by using the small surface-mount ACS780 or through-hole ACS770.

DESIGN WORLD — EE Network

Allegro - PE Handbook 2.17 V3 - update for digital edition.indd 38

2 • 2017

References ACS780LR data sheet www.allegromicro.com/en/ Products/Current-Sensor-ICs/ Fifty-To-Two-Hundred-Amp-Integrated-Conductor-Sensor-ICs/ACS780.aspx ACS770CB data sheet www.allegromicro.com/en/ Products/Current-Sensor-ICs/ Fifty-To-Two-Hundred-Amp-Integrated-Conductor-Sensor-ICs/ACS770.aspx

powerelectronictips.com | designworldonline.com

3/3/17 2:22 PM

POWER MODULES

Power modules lessen the load on internal combustion engines THOMAS YIM | ON S E M ICON DUCTOR

Integrated smart power modules will run mechanical loads that would otherwise sap fuel mileage.

GOVERNMENT REGULATIONS are stimulating demands for more economical vehicles that produce not just fewer NOx pollutants, but also less CO2. Vehicle tax policies currently favor less polluting models, and car makers must cope with new legislation such as the EU average fleet emissions directive which dictates the average CO2 emissions of all cars must not exceed 130 g CO2 /km. Non-complying manufacturers must pay an excess emissions premium on every unit sold. The CO2 limit will be reduced to 95 g/km by 2021 and to even lower levels as time goes on. An important part of the strategy for meeting such demands is to replace major subsystems usually driven by the engine with electrically-driven alternatives. Water pumps, oil pumps, air-conditioning units, turbo chargers and power steering systems have all been candidates. Reducing mechanical loads on the engine lets more fuel go toward powering the wheels. Replacing mechanical motor loads such as pumps, fans, compressors with an electric unit could reduce fuel consumption by as much as 3 to 5%. The automotive industry is keen to ensure that the replacement electrical systems are as efficient as possible. Another concern is how proliferating electrically-driven systems may grow the size and weight of vehicle wiring harnesses. Car makers must prevent harnesses from becoming excessively large and heavy. Additionally, electrical systems replacing mechanical functions will make heavy use of high-power motor drives. Here, the automotive industry can benefit from experience from other industries. In domestic appliance markets, for example, energy-efficient motor drives have been of paramount importance for years. Product-labeling schemes that let consumers compare energy ratings have focused attention on efficiency and enabled informed buying decisions. The cost of energy-saving variable-speed drives has come down with advances in power-semiconductor technologies and the maturing of intellectual property like Field-

powerelectronictips.com | designworldonline.com

OnSemi - PE Handbook 2.17 V1.indd 39

2 • 2017

The DBC (direct-bonded copper) substrate, as depicted on these chip package drawings, enables economical and efficient thermal management.

DESIGN WORLD — EE Network

2/21/17 3:17 PM

POWER ELECTRONICS

HANDBOOK

Oriented Control (FOC) firmware. Use of variable-speed drives has gradually filtered down from high-end equipment into mainstream products. One result has been appliances that cost less to run, are quieter, and are more versatile. Motor drives that are quiet, as well as highly efficient, are increasingly important for the automotive industry. As vehicles become more electric and adopt stop-start (micro-hybrid) operation to save fuel in combustion engines, vehicle cabins are becoming progressively quieter. Consequently, mechanisms like electric motors must also become more silent so they don’t A screen-grab of a FAM65V05DF1 power module in annoy passengers. operation illustrating the short circuit rate condition: Vdd=450 V,Vcc=15 V Tj=150⁰C. Three individual Variable-speed control isn’t just make electric motors more single-channel high-voltage ICs control the gates efficient. It also brings the opportunity to adjust oil and coolant of the high-side IGBTs. A single low-voltage IC with flow rates over a wide range of engine operating conditions. For three output channels controls the gates of the low-side IGBTs. example, suppose an engine has been running at high speed but then idles as the car waits in traffic. A conventional water pump will slow down when the engine does. In contrast, an electronically controlled variable-speed pump can be programmed to move coolant at a rate determined by the thermal conditions inside the engine. It can also continue running after the engine has switched off to optimize the engine cooldown profile. The main power circuit in a variable-speed drive is basically an inverter. And drives with inverters are more complex than simple fixed-speed drives. They comprise a controller that runs the FOC algorithm, high-side and low-side gate drivers for the power-transistor bridge driving current into the motor phases, up to six IGBTs or MOSFETs in a three-phase bridge, and protection against potentially destructive hazards such as voltage surges, short-circuit currents, and excessive temperature. Modern control modules for motor drives have a high level of integration that lets automakers save space, boost reliability, and keep down costs. And lessons learned with motor drives for the appliance industry have helped advance the trend. For example, Intelligent Power Modules (IPM) created originally for appliances take advantage of state-of-the-art fabrication and packaging technologies to combine high-voltage, high-power circuitry with logic circuitry all in the same device. An IPM integrates the high-voltage power silicon of the bridge in the same module as the gate drivers and protection circuitry. This integration has several advantages for motor-driven systems. For one thing, the modules have low thermal resistance because their circuitry sits on direct-bonded copper (DBC) substrates. Obviously, high integration also simplifies the design of the power stage – the circuit topology inside the IPM is not trivial.

DESIGN WORLD — EE Network

OnSemi - PE Handbook 2.17 V1.indd 40

2 • 2017

powerelectronictips.com | designworldonline.com

2/17/17 9:51 AM

It’s not a web page, it’s an industry information site Stay current with the latest electronic tips, resources, and news, visit eeworldonline.com and stay on Twitter, Google plus, Facebook and Linkedin. It’s updated regularly with relevant technical information and other significant news to the electrical design engineering community.

eeworldonline.com

EE Online house ad 2017.Vers2.LL.indd 41

2/21/17 1:09 PM

POWER ELECTRONICS

HANDBOOK

Moreover, it is space-efficient to concentrate all the driver power-stage circuitry in a single module. The module can occupy a smaller footprint and also typically will weigh less. Moreover, it has a high reliability because there are fewer individual components to be placed and connected. Recently, ON Semiconductor’s FAM65V05DF1 Auto IPM Smart Power Module became available. It is an automotive-qualified (AECQ100/101) integrated power module containing a threephase bridge comprising six power switches that each combine a 650-V high-efficiency IGBT and a freewheeling diode having soft recovery characteristics and a low reverse-recovery current. The IGBT’s 650-V rating allows operation from a bus voltage of up to about 400 V with adequate safety margin. This high-voltage capability is important because the industry increasingly contemplates using high-voltage buses to distribute power to electrical loads at low current. Three individual single-channel high-voltage ICs (HVICs) control the gates of the high-side IGBTs, and a single low-voltage IC (LVIC) with three output channels controls the gates of the low-side IGBTs. The LVIC and HVICs in the FAM65V05DF1 implement individual under-voltage lockout (UVLO) circuitry to protect the IGBTs against operation with too little gate-driving voltage. Over-current protection circuitry is implemented in the LVIC, with a soft turn-off feature that protects IGBTs from potentially damaging voltage surges by decaying the gate voltage instead of turning the devices off abruptly. The module’s temperature sensor, used to coordinate thermal protection, is also integrated in the LVIC. In addition, the LVIC has a fault output, which can be used to activate system-level protection for optimum reliability. For automotive use, module thermal performance is extremely important. Peak ambient temperatures can be high, especially for modules sitting near the engine. Size and cost constraints must also be taken into consideration. A DBC substrate is economical and thermally efficient. In the Auto IPM, the IGBT and freewheeling diode dies attach directly to the DBC substrate. This ensures efficient routing of heat away from the power dies to the edge of the package, where a heatsink can aid dissipation.

DESIGN WORLD — EE Network

OnSemi - PE Handbook 2.17 V1.indd 42

2 • 2017

The FAM65V05DF1 block diagram illustrates the integration of three HVICs, a three-channel LVIC and power semiconductors.

Bringing power semiconductors, drivers and protection circuitry together in a single module has enabled the realization of a complete threephase bridge and driver in a footprint of just 44 × 26 mm. This represents a 30% saving in PCB space compared to a conventional controller implemented using individual automotivequalified components (discrete IGBT + external three-phase gate driver, etc.). In a nutshell, energy-efficient integrated high-voltage power modules help eliminate power-sapping mechanical actuators typically driven by the engine crank. Field-stop-trench IGBTs, STEALTH diodes, HVIC, LVIC and DBC technologies are the major enablers for this new generation of auto IPMs.

References ON Semiconductor www.onsemi.com/PowerSolutions/home.do

powerelectronictips.com | designworldonline.com

2/17/17 9:55 AM

HOW TO BE SMART WHEN SELECTING A FAN

How to be smart when selecting a fan for forced air cooling JEF F SMO O T | CU I I NC.

It’s easy to over-specify a cooling fan. But there are ways of keeping temperature down with fans that are sized just right.

THE FIRST STEP in any thermal management task is understanding the thermal path taken for removing excess heat. Most systems, and particularly those that employ an enclosure, will use some form of forced air cooling. Invariably the cooling will involve a fan of some kind. The choice of fan, particularly for an enclosed PCB can have a significant impact on the overall lifespan of a system. Conduction is the simplest means to remove heat simply because heat spontaneously flows from a hotter to a colder body. However, an electronic assembly sitting in an enclosure such as a rackmount will not be able to dissipate much heat through conduction. For this reason, assemblies that consume as little as 25 W of power may require forced-air cooling. To grasp the scale of the cooling problem, engineers create what’s called a thermal profile. This is basically a record of how much heat a system generates, when it gets generated, and where the heat sources are. The approach to creating the profile starts with placing temperature sensors at various points on the PCB. Also important for system profiling is how much impedance to airflow a system exhibits. The system impedance is expressed in terms of a drop in air pressure between inlet and outlet. It plays a major part in calculating the overall airflow required from a fan and, in turn, the selection of the size and type of fan. Designers can determine system impedance by measuring the pressure drop using sensors or, if possible, by placing the

powerelectronictips.com | designworldonline.com

CUI - PE Handbook 2.17 V2.indd 43

An example of a computational fluid dynamics analysis (CFD) which can accurately profile cooling requirements

2 • 2017

DESIGN WORLD — EE Network

2/16/17 4:13 PM

POWER ELECTRONICS

HANDBOOK

system in an air chamber. For larger systems --such as data centers – it may be possible to model the system using CFD (computational fluid dynamics) to get an even more accurate profile. Forced air cooling plays a big role in thermal management, but the question becomes how much forced air is enough. To get the answer, it is necessary to figure out by how much the internal temperature can change without increasing the risk of failure. The process begins with determining the ‘most critical’ component in terms of operating temperature; this component’s maximum temperature will give a maximum ambient temperature allowable. The total power dissipation for all relevant components -- such as power transistors, microprocessors, amplifiers and communication interfaces -- will provide a figure for the amount of power the overall design dissipates. Power dissipated, in Watts, converts linearly to energy in Joules/second, which is in turn exhibited as heat. One can assume that the temperature of the air around the components will continue to rise while the equipment operates. At some point air temperature will reach a level that will inhibit the removal of additional heat. Thus he point of adding forced air is to replace the heated air with cooler ambient air. The air flow must be sufficient to accomplish this while keeping components below their maximum operating temperature.

Typical axial fan performance

A typical flow vs. pressure performance curve of an axial fan, this one for the CFM-120 series from CUI.

DESIGN WORLD — EE Network

CUI - PE Handbook 2.17 V2.indd 44

2 • 2017

Equation 1 shows the relationship between temperature rise and airflow, where q is the amount of heat absorbed by the air (W), w is the mass flow of air (kg/sec), Cp is the specific heat of air (J/ kg°K) and ΔT is the temperature rise of the air (°C): (1 )

Once we know the maximum permissible temperature within the enclosure and we derive the amount of heat generated (based on the cumulative power/heat dissipated by the components) it is possible to calculate how much airflow we need. Because mass flow (w) = air flow (Q) × density (ρ), substituting and solving for Q we can rewrite Equation 1 to get Equation 2 (where Q is the airflow in CMM (m3/min), q is the amount of heat to be dissipated (W) and ρ is the density of air (kg/m3)). (2 )

Finding the operating point

A typical performance curve of an axial fan with system impedance plotted overtop. The system operating point is where the two curves intersect.

powerelectronictips.com | designworldonline.com

2/21/17 3:19 PM

HOW TO BE SMART WHEN SELECTING A FAN

Substituting constants for Cp and ρ at 26°C, we can arrive at a general equation for calculating airflow, Equation 3. (3)

The calculated airflow figure can now be compared with the fan specification. Manufacturers characterize fans using these two parameters to provide a performance graph that plots airflow (measured in either CMM or in cubic feet per minute, CFM) against static pressure (measured in either inches or millimeters of water, often written as inch H2O or mm H2O). Consider the performance curve of the CFM-120 Series from CUI, a 120 ×120-mm frame axial fan with dual ball bearing construction. Unfortunately, the result given by Equation 3 is only accurate for ‘ideal’ conditions; with no back pressure from the enclosure (known as system impedance). In reality there will always be some system impedance. So to determine the real-world requirements engineers must calculate or estimate the system impedance. This impedance can then be plotted on the fan performance curve, and the point at which they cross should be taken as the operating point for the fan.

In many cases, it is not possible to measure the airflow through an enclosure using an airflow chamber. The alternative is to specify the operating point above the figure derived from Equation 3. For example, if the airflow calculated is 50 CFM with zero back pressure, there will be a good margin of error with a fan over-specified such that it produces a maximum of 100 CFM with the intention of operating it at 75 CFM. The higher figure also provides some headroom for increasing airflow during operation. Obviously, minimizing system impedance helps keep down the size of the necessary cooling fan. At a minimum, it is good practice to keep the areas around the air inlet and outlet as clear of components as possible. Also keep in mind that an air filter will add system

you could incorporate the switch functionality of a circuit breaker with the high protection level of a fuse?

New Fused Disconnect Switch UL98 Rated for CC fuses up to 30A & 600V The new Fused Disconnect Switch (FDS) series incorporates the switch functionality of a circuit breaker with the high protection level of a fuse. The FDS allows end-users to shut off and isolate branch circuits in electrical control systems in order to safely perform maintenance on the downstream circuit components. To view the product data sheet and learn more about the FDS, please visit: www.marathonsp.com/New Products/Fused Disconnect Switch Regal and Marathon are trademarks of Regal Beloit Corporation or one of its affiliated companies. ©2016 Regal Beloit Corporation, All Rights Reserved. MCAD16061E • SB0045E

CUI - PE Handbook 2.17 V2.indd 45

2/21/17 3:20 PM

POWER ELECTRONICS

Determining speed from sensor output

HANDBOOK

impedance. Component placement on the PCB should encourage airflow to and around critical components, using guides if needed. In addition, note that the above equations use air density at sea level. A system that will work at altitudes significantly above sea level must factor in the thinner air. A significantly higher altitude would need much more airflow to maintain the same level of cooling. A pulsed output from a tachometer can be used by control circuitry to gage and adjust fan speed.

Choosing the Right Fan Fans are generally categorized by the way the air enters and leaves. If it exits in the same plane as it enters then the fan is normally termed an axial fan, as to draw air in from one side and expel it from the other on the same axis. If the airflow leaves in a different plane the fan is normally referred to as a centrifugal design because the air drawn in changes direction inside the fan and is expelled in a different direction. This style of fan can effectively compress the air, allowing delivery of a constant airflow under different pressures. Perhaps the most prolific centrifugal fan design is the blower, which resembles an axial fan but typically expels air at 90° to the intake. The volume of airflow needed and the static pressure of the system will influence what style of fan makes sense for a given application. Axial fans are predominantly suitable for high airflow in systems with low static pressure, while centrifugal fans offer lower airflow but can deliver it against higher static pressure. Sensing stall/lock faults Both audible and electrical noise are also important when selecting a fan. The general rule of thumb is fans producing greater airflow will also generate more audible noise. Thus, axial fans will typically have lower audible noise than blowers. To keep down audible noise, designers must optimize airflow and reduce system impedance, thus reducing the required CFM. Besides generating audible noise, dc fans can have other design considerations. The dc motor within the fan creates an electromagnetic interference (EMI) signature. These emissions are normally limited to conducted EMI in the power leads. Ferrite beads, shielding or filtering can all mitigate such effects. For most PCB-based systems in an enclosure, the dc axial fan provides the right balance among cost, audible noise, EMI, and performance. Differences in the construction of axial fans may Fan can generate output signals indicating stall/lock fault. If the fan also impact how they are used. Generally, these motor stops, the output is driven to a logic high. It remains at a logic low during normal operation. differences relate to the bearings. Bearings on the fan shaft can be either steel ball bearings or sintered 46

DESIGN WORLD — EE Network

CUI - PE Handbook 2.17 V2.indd 46

2 • 2017

powerelectronictips.com | designworldonline.com

2/21/17 3:21 PM

HOW TO BE SMART WHEN SELECTING A FAN

PWM duty cycle

powdered bearings, usually referred to as sleeve bearings. At consistently low temperatures, fans with sleeve bearings can operate as well as ball bearing fans. However, at variable or high temperatures, ball bearings have been shown to operate longer and more reliably. Sleeve bearing fans, which are normally cheaper than ball bearing fans, do have their place. But their relatively short lifetime and propensity to fail at high temperatures limits their use. Axial fans are widely used in rack-mount enclosures thanks to their combination of small size, low power and high airflow. Many also include features that can boost system performance by controlling fan speed, optimizing a design to reduce power consumption. Design calculations of minimum required airflow rate help specify a fan that can cool under all conditions. They assume the fan runs constantly, even when maximum cooling is unnecessary. Thus the design calculation assumes worstcase conditions always. That means the fan is full-on even when cooling loads are not large. The result can be energy inefficiencies and shorter fan lifetimes. Consequently, it has become common practice to monitor the temperature in an enclosure and actuate the fan only when necessary. The downside is this practice can present problems in terms of thermal lag. It can also introduce a fault condition if, for some reason, the fan cannot start due to an obstruction. To address this problem, modern dc axial fans like the CFM Series from CUI include auto-restart protection as a standard feature. This feature detects when the fan motor cannot rotate and automatically cuts the drive current. Models including the CFM-60 Series also offer optional controls such as a tachometer and rotation detection sensors. The tachometer produces a pulsed output proportional to fan speed that can be used within control circuitry. If the motor stops, the output stops pulsing and stays at either a logic high or logic low. The rotation detection feature doubles as a lock sensor; if the fan motor stops, the output is driven to a logic high and remains at a logic low during normal operation. In addition, there is the ability to control the speed of the fan using pulse width modulation (PWM); the duty cycle of this input determines the speed of the fan rotation, the relationship between the duty cycle and whether the fan speed is linear. When used in conjunction with a simple algorithm

powerelectronictips.com | designworldonline.com

CUI - PE Handbook 2.17 V2.indd 47

running on a microcontroller it is possible to create a sophisticated thermal management scheme that is adaptable and efficient. A simple example of fan control could include temperature sensors distributed around a board. Each sensor could monitor a cooling zone, a particularly helpful concept when there are board components susceptible to heat variations. When the measured temperature approaches a predetermined level, the fan can turn on or speed up by changing the duty cycle of the PWM signal. Similarly, the fan can slow down if the internal temperature cools off. All in all, integrated circuits and PCBs are becoming ever more complex and dense. Statistically, most components fail because of overheating or operating too hot for too long. With so much to risk, use of the right fan can be the difference between a premature failure and an efficiently operating system.

References CUI Inc. www.cui.com

2 • 2017

DESIGN WORLD — EE Network

2/21/17 4:13 PM

POWER ELECTRONICS POWERELECTRONICTIPS.COM

Chroma Systems Solutions ............................................................... BC Coilcraft ........................................................................................... IBC Digi-Key ................................................................................ Cover, IFC Equipto .............................................................................................. 12 Fotofab .............................................................................................. 31 IDT ..................................................................................................... 25 IXYS ................................................................................................... 33 Keystone Electronic Corp. ................................................................... 1 Master Bond ...................................................................................... 18 Marathon Special Products ................................................................ 45 Mean Well ......................................................................................... 13 Memory Protection Devices, Inc. ......................................................... 3 PICO Electronics .................................................................................. 5 Rutronik ............................................................................................... 4 Sager Electronics.................................................................................. 7 Tamura Corporation of America ........................................................ 18

index LEADERSHIP TEAM

SALES

Publisher

EVP

Mike Caruso

Michelle Flando

Tom Lazar

Mike Emich

Marshall Matheson

mcaruso@wtwhmedia.com

mflando@wtwhmedia.com

tlazar@wtwhmedia.com

memich@wtwhmedia.com

mmatheson@wtwhmedia.

469.855.7344

408.701.7944

508.446.1823

com

440.670.4772 @mflando

@wtwh_memich

805.895.3609

Garrett Cona

@mmatheson

gcona@wtwhmedia.com

Mike Francesconi

Jim Powers

213.219.5663

mfrancesconi@wtwhmedia.com

jpowers@wtwhmedia.com

630.488.9029

312.925.7793

Managing Director Scott McCafferty

@wtwh_Tom

@jpowers_media

smccafferty@wtwhmedia.com

Jessica East

310.279.3844

jeast@wtwhmedia.com

David Geltman

@SMMcCafferty

330-319-1253

dgeltman@wtwhmedia.com

Courtney Seel

@wtwh_MsMedia

516.510.6514

cseel@wtwhmedia.com

@wtwh_david

440.523.1685

Michael Ference

@wtwh_CSeel

mference@wtwhmedia.com

Neel Gleason

408.769.1188

ngleason@wtwhmedia.com

@mrference

312.882.9867 @wtwh_ngleason

WTWH Media, LLC 6555 Carnegie Ave., Suite 300 Cleveland, OH 44103 Ph: 888.543.2447 FAX: 888.543.2447

DESIGN WORLD does not pass judgment on subjects of controversy nor enter into dispute with or between any individuals or organizations. DESIGN WORLD is also an independent forum for the expression of opinions relevant to industry issues. Letters to the editor and by-lined articles express the views of the author and not necessarily of the publisher or the publication. Every effort is made to provide accurate information; however, publisher assumes no responsibility for accuracy of submitted advertising and editorial information. Non-commissioned articles and news releases cannot be acknowledged. Unsolicited materials cannot be returned nor will this organization assume responsibility for their care. DESIGN WORLD does not endorse any products, programs or services of advertisers or editorial contributors. Copyright© 2017 by WTWH Media, LLC. No part of this publication may be reproduced in any form or by any means, electronic or mechanical, or by recording, or by any information storage or retrieval system, without written permission from the publisher. Subscription Rates: Free and controlled circulation to qualified subscribers. Non-qualified persons may subscribe at the following rates: U.S. and possessions: 1 year: $125; 2 years: $200; 3 years: $275; Canadian and foreign, 1 year: $195; only US funds are accepted. Single copies $15 each. Subscriptions are prepaid, and check or money orders only. Subscriber Services: To order a subscription or change your address, please email: designworld@halldata.com, or visit our web site at www.designworldonline.com POSTMASTER: Send address changes to: Design World, 6555 Carnegie Ave., Suite 300, Cleveland, OH 44103

DESIGN WORLD — EE Network

ad index — PT Guide 2-17 v1.indd 48

2 • 2017

powerelectronictips.com | designworldonline.com

2/21/17 11:44 AM

When it comes to power inductors, we’re spreading things a little thin (actual height)

In fact, with hundreds of options under 1.2 mm high, finding the perfect low-profile power inductor is no tall order! Let’s face it, thin is in. From smart phones and wearables to all types of portable devices, you face constant pressure to pack more performance into the thinnest packages possible. To help, we continue to expand our line of mini, low-profile power inductors with footprints as small as 1.14 x 0.635 mm and heights as low as 0.50 mm! Select inductance values from 0.018 to

3300 µH and current ratings up to 20 Amps. Get the skinny on all our low-profile power inductors, including the new ultra-low loss XEL4012 Series with inductance values that have been fully optimized for high frequency applications over 5 MHz. Visit coilcraft.com/lowprofile today! ®

WWW.COILCRAFT.COM

Coilcraft PEH 2-17.indd 77

2/16/17 4:32 PM

Simulate DC Transients 62000P Programmable DC Power Supplies

Program DC transient waveforms to simulate automotive and defense voltage variations

Optional Softpanels

Chroma’s 62012P-80-60 Conventional power supply

Chroma’s 62000P series of programmable DC power supplies allows devices to be verified under a wide range of voltage dropouts, spikes and other voltage variations making the P-series an ideal choice for production and DVT of products used on mobile platforms or which will experience voltage interrupts. Applications include products designed for use on ships, cars, planes and other applications with variable or unstable input power. Due to their constant power operating envelope the P-series power supplies are uniquely suited for test applications requiring a wide range of high voltage/low current and low voltage/high current stimulus thereby reducing the number of supplies needed in typical applications. The 62000P Series also includes built in 16 bit readback capability for accurate input voltage and current readings. This means ATE systems no longer need complex shunt/multiplexers to make accurate pass/fail readings of the UUT’s input parameters. These instruments also include I/O ports providing 8 bit TTLs, DC-ON, fault output signal and remote inhibit as well as an output trigger signal for system timing measurements. 12 different models ranging from 600W to 5000W, up to 120A and up to 600V.

Wider Operating Region

For more information visit: chromausa.com

Instruments AC Power Sources Regenerative Grid Simulators Programmable DC Power Supplies AC & DC Electronic Loads Power Meters Multimeters

Automated Test Systems Hipot Testers and Analyzers Wound Component Testers LCR Meters Milliohm Meters TEC Controllers Thermal Data Loggers

Battery EV/EVSE PV Inverter Power Conversion Medical Device LED Lighting and Driver

Chroma System_1-17.indd 165

2/16/17 4:33 PM