Sursa proiect.docx

8/19/2019 Sursa proiect.docx

1/18

Elliott Sound Products Project 89

Switchmode Power Supply For Car AudioSergio Sánchez Moreno and Rod Elliott

Share |

Foreword

This contributed project is a result of considerable collaboration between Sergio andmyself, and should not be seen as necessarily a complete project in itself, but astepping stone to understanding switching power supplies, how they wor, and whatyou can do with them!

"e warned # there is considerable ris! "ecause of the e$tremely high currenta%ailable from a car battery, a tiny mistae may easily lead to catastrophic failure! &llelectronic components are said to contain smoe 'wire contains an enormousamount(, and a slip of the soldering iron can liberate an unbelie%able )uantity!Seriously though, the ris of se%ere burns and the possibility of causing a fire in yourcar are %ery real, and should not be underestimated! *++& from a car battery can do

a %ast amount of damage in a few milliseconds # should the fuse not blow 'youwill

use a fuse, wont you-(, then the damage can be e$tensi%e!

&t %arious points in Sergios part of the article, . ha%e included some additionalinformation!

Please see the special note at the end of this article for important information about theproject.

Introduction

The difficulties of installing a hi#fi system in a car are many, although there is nodoubt that the most important is the limitation of the %ehicle supply %oltage! &s mostreaders already now, the nominal %oltage of a car battery is /01, reaching about/*!21 when charging 'i!e! engine running(!

The ma$imum RMS! audio power from a gi%en %oltage 1 is somewhat less than3

4ma$ 5 ' 1 6 ' 0 $ 70 ( (8 6 R9

!!! where R9 is the speaer nominal impedance!

Thus, for a /*!21 system, this power is limited to about :; on a < =hm load! >otethat the lower the resistance of the speaer, the higher the ma$imum power 'this is

http://addthis.com/bookmark.php?v=250&username=rodehttp://www.sound.westhost.com/project89.htm#specialnotehttp://www.sound.westhost.com/project89.htm#specialnotehttp://www.sound.westhost.com/project89.htm#specialnotehttp://www.sound.westhost.com/project89.htm#specialnotehttp://addthis.com/bookmark.php?v=250&username=rode


2/18

the reason most audio speaers ha%e a < ohm nominal impedance instead of themore common 2 ohm in home systems(!

This may be simplified to some e$tent !!!

4 5 ' 1 6 * (8 6 R9

and a typical calculation based on a /*!21 supply gi%es !!!

4 5 ' /*!2 6 * (8 6 <4 5

This allows for standard losses, and is acceptably accurate at this %oltage # the onlyreal way to now is to measure the amp, since the losses %ary depending on thetopology of the output stage!

4ower output can be increased by a factor of nearly < by using bridging techni)ues,e$plained in more detail in ES4 project /


3/18

M=SETs or high speed bipolar transistors are used 'some SM4S designsuse SBRs but these are in the minority(!

• =nce this wa%eform is stepped#up by the transformer, it has to be rectifiedagain and filtered bac to AB, since that is what we want! or audio

applications, we usually need a symmetrical supply, F6#*?1, for e$ample! Therectification is done with a diode bridge, as it would be using a con%entionaltransformer at ?+ or :+ Cz! >ote that for the fre)uencies we are taling about,fast or ultra#fast diodes are needed!

• .f we need a regulated power supply, some ind of feedbac must be pro%idedfrom the output rails to a controller that can change some parameters of the &B wa%eform at the primary of the transformer! This is normally accomplishedwith 4;M 'pulse width modulation(! ;e will e$plain this later, in theregulation paragraph!

• &lways eep in mind that no energy is created # gi%en a 'total( rails to battery%oltages ratio, the current drawn from the output will be 'at least( be multipliedat the /01 input by the same ratio, thus the total power stays the same'assuming /++G efficiency, and that is ne%er the case(! & generic transformertransforms the %oltage by a factor of Tr, current by a factor of /6Tr, andimpedance at the secondary by a factor of / 6 7'Tr(, Tr being the turns ratio!.mpedance is of little importance in this conte$t!

• & well built SM4S can reach @+G efficiency! So, if you e$pect to produce H*?1at :& 'per rail( supply 'this supposes *?$: F *?$:5*:+;( then be prepared to

draw more than *+& from the batteryI ortunately, when taling about audioamps reproducing music, power re)uirements are always much lower thanwith pure sine wa%es!

&t this point, the reader should realise the magnitude of the currents in%ol%ed in ahigh power SM4S for a car amplifier, and that e$treme caution should be taenespecially when connecting the creature to the car electrical system!

"he system

The present project describes the construction of a fle$ible SM4S capable ofdeli%ering powers in the order of *?+; continuously, depending on the transformerused! The output %oltage depends mainly on the turns ratio of the primary andsecondary windings, but may be adjusted to a somewhat lower %alue usingregulation! This should be enough to power a 0++; subwoofer amplifier plusperhaps 0 stereo amps for the mids and highs!

.t is part of a complete car amp that . ha%e built, with : power stages based on>ationals 9M*22: =%erture &mplifier! They can be combined into one J0?+;6< =hmsubwoofer channel plus 0 $ :?;6< =hm midFhigh channels, alternati%ely into 0 $/0+;6< =hm F 0 $ :?; or e%en to form a multichannel : $ :?;6< =hm amplifier, soit is an e$tremely fle$ible and high#powered system without renouncing sound )uality!


4/18

The parallel bridging techni)ues needed to do this will be possibly described inanother project!

Construction of the SPS

The complete schematic of the SM4S is shown below!

#ote$ "his is Ser%io&s ori%inal 'ersion of the supply( the one shown in Fi%ure 9 is li)ely to bethe most commonly used( as it is somewhat simpler( but has 'irtually identical performance.*esp+

Fi%ure , - Switchmode Controller Schematic

There are three main blocs described below !!!

http://www.sound.westhost.com/project89.htm#fig9http://www.sound.westhost.com/project89.htm#fig9


5/18

& # Switching M=SETs and transformer" # Rectification and filteringB # Bontrol circuitry

A - Switchin% SFE"s and "ransformer

The selected switching topology is called a push#pull con%erter, because thetransformer has a double primary 'or a centre#tapped one, if your prefer(! The centretap is permanently connected to the car battery '%ia an 9B filter to a%oid creatingpeas in the battery lines, which could affect other electronic e)uipment in the car(!The two ends of the primary are connected to a pair of paralleled M=SETs eachthat tie them to ground in each conduction cycle '1gs of the corresponding M=SEThigh(!

These M=SETs should be fast, able to withstand high currents 'in e$cess of *+&each if possible( and ha%e the lowest possible Rds'on(! The proposed =n#Semiconductors MT4K?>+: can withstand K?&mp and has a Rds'on( below /+

milliohm! This is important, because the lower this resistance is, the less power theyare going to dissipate when switching with a s)uare wa%eform! &nother alternati%esare MT4:+>+:, or the more popular "DL// and .R?


6/18

The resistor RT and capacitor BT fi$ the oscillation fre)uency! E$perimentationshowed me that about *?Cz produces good results with my transformer! &nothercapacitor, Bss fi$es the soft start time # when you turn on the system, the pulsewidth increases from + up to the steady %alue, thus limiting the inrush current, a %erygood feature to a%oid thumps in the speaer and protect the electrical installation! .t

has also a shutdown pin that allows control of the SM4S from an e$ternal signal'REM=TE from the head unit, for e$ample(!

.n this project, layout is critical, incorrect trac widths or e$cessi%ely long traces canha%e high inductances and produce peas that can mae the M=SETs blow up!

"ransformer Construction 0etails

This is the most critical part of the design, and you ha%e two options, buying acommercial unit with the re)uired power rating and turns ratio 'hard to find, only a

single supplier found at the time of writing(, or wind your own!

.f you choose to wind your own transformer 'as if you ha%e much choice(, you ha%eto decide which shape of core to use! The preferred material is ferrite, which has highpermeability 'ability to conduct magnetic flu$( or iron powder, which has a lowerpermeability, but is less liely to saturate! Most commercial transformers use ferrite,and iron powder is generally the best material for filter choes 'inductors( that carrysubstantial AB!

or e$ample, with a standard ETA*@ core you could theoretically build a J *?+;supply! ;inding this type of cores is not %ery difficult, but you will ha%e to follow someguidelines . pro%ide below in order to ha%e good results!

&nother possibility is using a toroid! Nou can e$tract it from a ". power inductor! &sa guide, a


7/18

E"0-type cores

"oroid from I"1 ,22 inductor 34ilco Corp5.3/emo'e the thic) wire before windin%65

These are a few general winding guidelines for all types of cores3

• Nou MDST use enamelled copper wire for all the windings! Oeep also in mindthat when woring with high fre)uencies, the effecti%e section of the wire ismuch smaller than the physical one, due to the sin effect 'the currentconcentrates only in the outer part of the wire(! &s high currents are in%ol%edhere, the section of the wire is %ery important, 'if you dont want the enamel tofuse due to the heating produced by the resisti%e losses of the wire and shortall the windings(! & good practice is to use se%eral thinner wires in parallelrather than a single thic one! This also eases winding! or e$ample, si$+! TCE S&ME A.REBT.=>(! To calculate the numberof turns of the secondary winding, multiply by the turns ratio! or e$ample, ifyou want to build a H*+1 supply, the turns ratio is *+6/*!250!0 appro$, so wind0!0 $ < 5 2!2 turns 'you cant ha%e a partial turn, so use @ turns( for each


8/18

secondary 'that is, @ turns, centre tap and another @ turns .> TCE S&MEA.REBT.=>(!

• To start winding, tae the number of thin wires you ha%e decided to use ':, fore$ample( in the primary, all together! 9ea%e about *+ or TCE S&ME A.REBT.=>winding the other < turns and at the end lea%e another *+ or


9/18


10/18

y system&s input cho)e( obtained from an old PC power supply.

• &ll the wiring, especially the primary side must be hea%y gauged, in order tominimise losses and a%oid o%er#heating of the conductors! The 4B" tracsshould be thic enough, as short as possible, and reinforced with a generoustin layer and possibly with soldered wire!

• 4ut two fuses in the rails outputs, as they can sa%e you a lot of headaches

when you short them to ground, etc! . used two standard :!*& fuses!

• Mount the rectifier diodes and the M=SETs on a decent heatsin, and eepin mind that they must all be electrically isolated! ollow the usual heatsinmounting recommendations 'thermal grease, etc!(! T=00+ pacages are easyto handle, but their thermal performance is mediocre so considerable care isneeded with mounting!


11/18

0etail of the SFE" arran%ement

>ote the insulation pad 'one for all( and the thic supply wires! .ndi%idual insulationpads may be used with no loss of performance! Dse of a clamping bar will gi%eimpro%ed thermal conduction to the heatsin bracet, but do not o%er tighten, or thebracet will bend! Bonsider using a bracet of not less than *mm aluminium for better thermal performance! The bracet must be attached to a heatsin, using thermalcompound to minimise the thermal resistance!

"ests

This project handles )uite large powers, so it is well worth the pain of step#by#steptesting before you regret blowing all your wor up in a microsecond!

or the tests, use a big /01 to /*!21 power supply, with current limiting if possibleand capable of deli%ering at least /+ to 0+ amperes 'see project KK(! .f you dont ha%ethat, a 4B computer 4SD will wor 'although you wont get more than 2+#@+;, but itis enough for testing purposes and almost indestructible(! Aont connect the SM4S toa car battery the first time you test it 'it can be really dangerousI(! & /+& fuse inseries with the /01 supply is also a good idea! 'Nou dont now to what e$tentI (

The cables from the supply to the amp should be as short as possible and hea%ygauged, to minimise losses! irst time . tested the amp . had a / %olt of differencefrom one side to the cable to the other in only /!? metres3 the cable itself wasdissipating more than /?;III! So, when calculating efficiency, always measure input%oltage just at the input of the SM4S to account for this!

The wires from the transformer to the M=SETs must be as short as possible! E%ery/+mm of wire adds around :nC of inductance thats e$ternal to the transformer 'thee$act figure depends on the wire diameter and can only e%er be an estimate(! This,and transformer leaage inductance, will cause %oltage o%ershoot and ringing on the

switching wa%eform! .f at all possible, the M=SETs should be no more than 0+mmfrom the transformer!


12/18

• irst of all, with only the S*?0? chip and its associated components 'noM=SETs(, chec that you ha%e a %ery clean /01 s)uare wa%e at each output'/2+P out of phase and they do not o%erlap E1ER(! Bhec also that when youturn#on the power, it starts from +G and increases to ?+G duty cycle in abouta second or two!

• =nce you ha%e this, you can wire the M=SETs! They will be on a heatsinand be aware that the tabs are connected to the drain, so pro%ide insulation'mica F plastic washers, thermal compound(! Then solder the transformer andwatch the primary wa%eform with an oscilloscope 'use a /+3/ probe just incase you ha%e large spies in order to a%oid damaging the instrument(! Noushould ha%e a s)uare wa%e of about 0?#0:1 pea to pea and the smallestpeas 'o%ershoot( possible! .t they are higher than *+1 'from ground(, youmay try to re#wind the transformer to impro%e coupling! Nou can also reducethe o%ershoots using the snubber networ shown in the schematic, althoughthey will dissipate a bit of power 'use 0; resistors and /++1 capacitors(, somount them only if necessary!

• =nce you ha%e a clean wa%eform, you can solder the rectifier and outputcapacitors and see what you ha%e in the positi%e and negati%e rails! Noushould ha%e the same %oltage in both, and it should be similar to what youcalculated!

• >ow load the supply with power resistors! Start with low power consumption'about 0+;( and obser%e the M=SETs, rectifiers and transformer carefully tosee that they dont heat up! &lso watch the current drawn from the /01 supply!

The power '1 $ .( should be only a bit higher than that at the output load!'E$pect a 2+G efficiency or so(!

• .f e%erything goes well, increase the load 'decrease its resistance %alue(! TheM=SETs should get warm after a while with hea%y loads 'about /++;(, andthe efficiency should maintain high 'always abo%e K?#2+G(!

;hen you are completely sure that e%erything wors as e$pected, you can proceedto connect it to the car electrical wiring 'see installation procedures paragraph(! irsttime you will notice an spar due to the sudden charge of the big input capacitor,unless you connect a resistor in series first '%ery good practice( to allow it chargingslowly and then remo%e it for normal operation!

Installation procedures

or your car and own safety, it is 1ERN .M4=RT&>T that you pay special attentionwhen installing the power supply 'and amplifier( in your car! These are somerecommendations that e%eryone should follow carefully3

• The supply MDST be taen directly from the battery, not to the radio or other

F/01 cables, as you will just blow or burn them, with the ris of a fire in the


13/18

car! The supply wire must be of ade)uate section, at least ? mm diameter'e$cluding the plastic co%er(!

• & fuse MDST be connected in series with the supply wire, as near the batteryas possible! .n case of a collision, the wire can be shorted to the chassis which

;.99 cause a fire! This is not a joeI The battery can pro%ide in e$cess of *++ & that can burn %irtually anything in a fraction of a second!

• The .RST connection you ha%e to mae to the amp is round, and that mustbe firmly screwed to the car chassis as near the amp as possible with thicwire! >otice that if you connected, for e$ample, the signal RB& cables first andthen the F/01 wire, the input capacitors would try to charge returning toground %ia the audio cables, possibly damaging the preamplifier of the headunit!

/e%ulatin% the Power Supply

The project itself has e$cellent load regulation, and the rails %oltage is almost onlydetermined by the turns ratio, but it has inherently zero line regulation 'basically, itsimply multiplies the input %oltage by the turns ratio(, although this is not a problemin a car where the battery %oltage remains essentially constant!

.f the obtained output %oltages are %ery high and you cant 'or dont want to( modifythe windings, you can use regulation to lower them a bit! or e$ample, . use a *3/transformer that would gi%e about F6#*21 without regulation that is unacceptable for

my 9M*22: stages to be safe, so . ha%e regulated to F6#0:1! The M=SETs willsuffer more, howe%er, so regulate the supply only if strictly necessary!

Nou can install the feedbac potentiometer and set it in order to ha%e zero reference%oltage to deacti%ate regulation, or increase its %alue to regulate to the desired%oltage!

>=TE3 Regulation will wor better with output inductors just between the rectifierdiodes and the output capacitors! /+ to /++QC with iron powder core and at least 2&current rating will be ade)uate! '. dont use them and my supply wors reliably,although . ne%er put it to the power limits(! Nou can also impro%e safety by paralleling

more M=SETs, so the current through them is shared! This also impro%es efficiencya bit, as the total Rds'on( is reduced!

btainin% -,:; from the SPS for Preamplifiers

.f you need to power opamps for a crosso%er, e)ualiser or preamplifier, you canobtain a symmetrical F6#/01 'for e$ample( from the main supply rails, simply with aresistor, zener and capacitor! 'see igure / of 4roject 0K(! Remember to use / or 0;resistors and zener diodes! Nou can obtain about 0?#?+ m& from this withoutproblems!


14/18

Additional Information

The following material is from ES4 # there are some suggestions and additionalinformation, as well as a simplified %ersion of the SM4S!

&lthough my %ersion of the switcher is simplified, this does not imply thatperformance is lower than Sergios original, but is the result of my own e$perimentsand tests! ;e may be on opposite sides of the planet, but there was considerablecollaboration during the de%elopment of the supply, and . ha%e built and tested the%ersion shown below!

SFE"s and "hermal /unaway

.t has been claimed that M=SETs are immune from thermal runaway, since theyha%e a positi%e temperature coefficient for their on resistance! ;hile this may be

partially true for a Blass#&" power amplifier, it is completely false for a switchingsupply!

or e$ample, a push pull SM4S using one .R?


15/18

. strongly recommend that you do the calculations yourself, and mae sure that youunderstand the implications!

/e%ulation

>ormally, one would e$pect regulation as shown in igure /, howe%er, using thefeedbac input of the controller .B relies rather too hea%ily on the impedance of theAB supply lines! >ormally, output inductors are used 'with an additional flybacdiode( to pro%ide a pulse width to %oltage con%erter! The majority of commercialsystems seem to use a non#regulated con%erter, so . would consider that this will be)uite acceptable in practice! Tests so far ha%e shown that with a load of about /?+;atts, the regulation was almost entirely dependent on the %oltage drop in the supplylineI

&s well as being unregulated, there are a couple of other changes in the circuit! R2

'/++ =hms( is connected between the timing capacitor and discharge pins of thecontroller .B! This introduces a dead time where both outputs are turned off, and thereason for this is to ensure that the power M=SET pairs can ne%er be switched onat the same time # should this happen, a %ery large current will flow 'albeit for only amicrosecond or less(! Since . did not use the e$tra switching transistors and usedhigher %alue gate resistors, the dead time is important!

. also increased the switching fre)uency! &s shown, the internal oscillator runs atappro$imately ?+Cz 'my prototype actually runs at ?ot surprisingly, the turns ratio is %ery important if regulation is not used! &ssume aninput %oltage of /01 to allow for losses! To obtain F6#0


16/18

Fi%ure 9 - Simplified ;ersion of Switchin% Supply

Since the transformer is relati%ely easy to wind, it is not a difficult tas to dismantle itand add 'or remo%e( secondary turns to get the %oltage right! My prototypetransformer used ?F? turns for the primary, and . used * strands of +!2mm windingwire twisted together! There is plenty of room in the recommended core, so it wouldbe easy to use ? strands instead for lower losses!

>ote that in the abo%e 'ig! @(, the hea%y leads shown carry substantial current, andmust be sized accordingly! . do not recommend 4B" traces be used, since thecurrent in%ol%ed is simply too high! i%en that the suggested current density for 4B"tracs is

. also eliminated the relay, but at the cost of a small current when the unit is notoperating! The S*?0? has a shutdown pin for just this purpose! & signal from theremote head amp will turn on /, and remo%e the shut down signal from thecontroller! .t beha%es in e$actly the same manner as if power had just been applied,

and the unit will become fully operation in about 0 seconds or less! Burrent drainwhen turned off will be about / to 0m& # considerably less than the cloc in the car!


17/18

"attery discharge will not occur as a result of this %ery small current, which may beignored as insignificant! eel free to use the relay if you prefer, connected as shownin igure /!

Construction

. recommend that an EAT*@ ferrite core is used! These are easy to wind, and arecapable of around *?+; output! "ear in mind that this represents a considerablebattery current at full power, in the order of *+ to *? &mperesI Cea%y transformerwindings and supply cables are essential, and the input filter must be capable ofwithstanding this current without saturating the core!

The former for these cores is rather large, and you may decide to cut the mountingsections off completely! Ao remember that the transformer must be mountedsomehow though, so . suggest that you ha%e a plan! &t this stage, . ha%e only

e$perimented, and do not ha%e a plan!

&ll of Sergios pre%ious comments apply to this %ersion, so mae sure that you readhis material thoroughly! . do not propose to co%er the same instructions again, sinceSergio has already done an e$cellent job!

Prototype "estin%

. ha%e done some initial tests, but ha%e not yet connected the bridge and outputcapacitors! ;ith what was intended to be /0F/0 turns on the secondary, . obtainedan acceptably clean wa%eform with some o%ershoot with the secondary unloaded!=utput %oltage was about *21 pea, so . ob%iously had one more turn than . thought. did 'input %oltage was /


18/18

This project has already created far more )uestions %ia e#mail than . desired ore$pected! or e%eryone who plans on maing this supply !!! you are essentially onyour own! . cannot 'and will not ( be drawn into lengthy e#mail e$changes if youcannot mae the supply wor!

That it does wor if built as described is certain, that you will be able to achie%e thesame results is not! .f you do not ha%e 'or at least ha%e access to( an oscilloscope #dont e%en thin) about trying to mae the supply, as it will not be possible to ensurethat the duty cycle of the controller is e$actly ?+G, or that there is no se%ereo%ershoot or ringing at the output!

4lease do not not send me e#mails asing for help! . will simply refer you to thisparagraph # . cannot diagnose your problems %ia mail, and will not e%en try! .t isentirely up to the constructor to determine his6 her abilities before starting!

The construction of any switching supply is fraught with difficulties, riss 'including

but not limited to electrocutionI( and problems that need to be addressed! They arenot simple 'despite appearances( or easy, and there are a great many things that cango wrong! .f you are not /++G confident that you understand the issues in%ol%ed,please do yourself a fa%our and build something else instead!

Projects Inde

Sursa proiect.docx

Documents

Transcript of Sursa proiect.docx