Web: http://www.pearl-hifi.com 86008, 2106 33 Ave. SW, Calgary, AB; CAN T2T 1Z6 E-mail: [email protected] Ph: +.1.403.244.4434 Fx: +.1.403.245.4456 Inc. Perkins Electro-Acoustic Research Lab, Inc. ❦ Engineering and Intuition Serving the Soul of Music

Please note that the links in the PEARL logotype above are “live” and can be used to direct your web browser to our site or to open an e-mail message window addressed to ourselves. To view our item listings on eBay, click here. To see the feedback we have left for our customers, click here.

This document has been prepared as a public service . Any and all trademarks and logotypes used herein are the property of their owners. It is our intent to provide this document in accordance with the stipulations with respect to “fair use” as delineated in Copyrights - Chapter 1: Subject Matter and Scope of Copyright; Sec. 107. Limitations on exclusive rights: Fair Use. Public access to copy of this document is provided on the website of Cornell Law School at http://www4.law.cornell.edu/uscode/17/107.html and is here reproduced below:

Sec. 107. - Limitations on exclusive rights: Fair Use

Notwithstanding the provisions of sections 106 and 106A, the fair use of a copyrighted work, includ- ing such use by reproduction in copies or phono records or by any other means specified by that section, for purposes such as criticism, comment, news reporting, teaching (including multiple copies for class- room use), scholarship, or research, is not an infringement of copyright. In determining whether the use made of a work in any particular case is a fair use the factors to be considered shall include:

1 - the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;

2 - the nature of the copyrighted work;

3 - the amount and substantiality of the portion used in relation to the copy- righted work as a whole; and

4 - the effect of the use upon the potential market for or value of the copy- righted work.

The fact that a work is unpublished shall not itself bar a finding of fair use if such finding is made upon consideration of all the above factors

♦ PDF Cover Page ♦ ♦ Verso Filler Page ♦ IMPROVING the WILLIAMSON AMPLIFIER

By T ALBOT M. WRIGHT Changing a few resistors in the driver section results in reduced distortion in this hi-fi amplifier design.

HE Williamson amplifier, whether it clium to high levels and instability. We'll A careful inspection of any U.S. tube be of the classic triode type or the discuss instability later-but first to the manual will disclose the fact that the T "Ultra-Linear" vel-sion, has to some distortion. 6SN7 or 6CG7, as used in the William­ extent fallen into disfavor with audio­ son amplifier circuit, is badly under­ IM Distortion philes since it is now possible to build an biased. amplit\er at lower cost and still equal or If you run an intermodulation distor­ Take the push-pull drivel' stage fO/' better the Williamson's performance. tion test on an avemge Williamson am­ instance. Usually the plate loads arc Many Williamson owners have dis­ plifier using the classic drive system you 47,000 ohms and the cathode resistor is carded their trusty old amplifiers in will find that the 1M exceeds 2'ir at only around 560 ohms. This produces a bias favor of 50 or 60 watts in the never­ slightly over half the rated power. By of around 51,2 volts with about 175 volts ending chase after unmeasurable dis­ the time you reach rated power you on the plates, assuming a supply of tortion. Many Williamsons have been find the distortion is well past the toler­ around 440 volts. The tube manuals. in­ converted to higher powel's by the use able level. The result of this testing dicate that the 6SN7 should be operated of newer tubes and heftier output trans­ usually leads to a substitution of all the with 250 volts on the plate and an 81J.a­ formers. In some cases the proud owner tubes in a wistful search for lower dis­ volt bias. has realized an audible increase in defi­ tortion. Sometimes this helps a little. All you have to do is change that nition and an accompanying drop in Basically {ne voltage amplifiers are cathode bias resistor from its present distortion. producing much greater amounts of dis­ value to 1000 ohms. A resistor of 1200 Frequently such an experimenter will tortion than they should. "But," you say ohms works well, too. With a lOOO-ohm add a high-power output stage to his to yourself, "I followed the design bias resistor, there .will be about 250 Williamson drive system-in most cases scrupulously; even the voltages are volts at the driver plates and the bias paying for 50 watts in order to get a what they should he." The point over­ will rise to approximately 9 volts. Not clean 20 watts. Even then he doesn't looked. however, is that the American only does this cut distortion but it always get the improvement he expects. 6SN7, 6CG7, or 12AU7 will not work means you can produce more drive volt­ The fault lies not in the power stage but well with the circuit values specified by age. Naturally, too, you can now put in the drive system. The two main faults Mr. Williamson in his original paper on more signal voltage into the 6SN7 driver of the Williamson are distortion at me- the subject. before it starts to draw grid current. The driver will stay class A, in fact long after the output stage has finally given Fig. 1. aeductlon In Intermodulatlon distortion of the amplifier shown in Fig. 2. up. I rhis change alone, of course. will not 14 eliminate the distortion fed into the out­ IJ put stage. The first two stages are still " 8£FO"£ underbiased and they hang together be­ It cause of the d.c. coupling between them. / Let's start with the first stage. The I 2' bias resistor usually found there is 470 I- 'I 11:10 ohms. This should be increased to twice o I­ its value-simply add another 470-ohm ell • / o / I resistor in series with the fil'st. This ' raises the bias to around 3 volts, but 2 I incl'easing this bias causes the plate � 7 voltage of this stage to go up, thereby I AFTER .1 • decreasing the grid-cathode bias aCI'oss 'I 'I the' phase-inverter-the next stage. i/ J There al'e now two things to be done: 4 lower the supply voltage to the first / / stage and increase the supply voltage 3 ./ to the phase-inverter. If each of these is t changed in the right proportion, both V stages will be biased correctly...... / , The average WiIliamson circuit uses a 33,OOO-ohm resistor from the 450-volt 4 I , 10 IZ 14 " I, to Zt Z4 ZI ze 30 3Z 34 3' 38 40 4Z 44 4, supply point to decouple the first stage; POWER-WATTS If0U4V. SINl-wAVE) this must be increased to about 47,000

104 ELECTRONICS WORLD ohlll�. This will cut the current through With a 450-volt supply you can get well the distortion should be well below 0.5";' . the stage somewhat, producing a cath­ over 100 volts to each gdd. With a 400- Performance of the amplifier of Fig, 2 nde uias of around 21, .. volts, and it will volt supply that drops to something near is illustrated in the curve of Fig. 1. For also lower the plate voltage. The de­ 350 volts due to lack of regulation in the all IM tests, the generator supplied fre­ coupling resistor to the phase-inverter power supply, you should still be able to quencies of 60 and 6000 cps at a ratio of is usually 22,000 ohms from the 450-volt get 80 volts of drive per grid. 4 to 1. Readers should not be too con­ point. This resistor should be l'ec1uced to There is one thing more which must cemed about the absolute values of 3900 ohms, pulling more current be done before you have the amplifier power indicated nor a comparison of the through the phase-invertel' and raising operating correctly. The output trans­ power in one amplifier we have worked the cathode uias, formers used in Williamson amplifi ers on with another. The important thing \\'hen these last two chan ges have will not usually hane-lIe more than about heing shown here is the definite im­ b(,l'1l made there should be about a 5 %­ 20 db of feedback, but increasing t.he I'e­ provement in performance that occurred to i-volt bias from grid to cathode of sistor to the cathode of the first stage with the changes incorporated. t Il!' phase-in\'el'ter. Of couI·se. if t he ini­ has already increased the feedhack. Amplifier No. 2 (see Fig. 3) uses four tial supply voltag\? is much below 450 Since the value of the cathode resistor 5881's in a push-pull parallt'1 triode out­ you will not get exact Iy the ri ght bias, has been doubled, it is necessary to put stage. In this one only the biasing but it will be close enough to allow the double the value of the feedback l'esistor modifications have been made, The /irst two stages to bias each othel' cor­ to put the feedback at its previous value. tubes were chosen at random and no at­ rectly. Typical volt ages for this circuit With the changes outlined, any "UI­ tempt was made to balance the drives are found in Fig, 2. tra-Linear"-type Williamson should not beyond the tolerances provided by the The net result of all these changes is produce 1 ';', distortion until it reaches 5-year-old 5"/� resistors already in the to almost double the drive voltage avail­ overload. If the drive volt ages to the circuit, A 450-volt capacitor-input able to the grids of the output tubes. grids of the output tubes are balanced, power supply was used in this circuit.

Fig, 2. Circuit of one of the amplifier, modified by the author. The value, of five re,i.tors have been changed, cs

la 1\

... ..

* CHI'NG£O COMPON£NTS

It ,._-/ ,"t'toA,.. ,,,,4;0 I.,,., '01 ...._------.... It :.R .--I.S "." "Ie,... I., .,. " ,. N :� . 2SfI u""" 10 1Il'•.. ;',rro.,,1I ,"s. C" 61111 Il/.• ISII ... . Iu. «,.,·i/o. R:.R;- IlIll(Jn flIe'H. �.', ... ,,.•. N .. :- UII.I/I/l1 oItHt. I .,. .... (.'... . (.' ::. (.' : •. C..:-'O Il/. . SIIII1I. r/u. ,.,.,·iI .. , 1I,-U,OI/(/ dm, '/, ... . N ...-- /1/11/0 d/H. I ... 'rI. J 1 /'''",'0 j.ri It.- ·'111 dHt. 1 ... 'rI. 1: •• (.'_-.1 Il/.. 11111. .. ,e,.,·i/o. (.'/1. 1 Ay .• .HI INe •• SIIII oA." rAo". 11•• 11•• 11 ".N.,-41.III/O 04 .... 1 ....,. . (.' .- - IS Illl/•• 11/11 11. c.,,,·i/o. C/I:.(.'II.-6 Ay .. 611·/110 d",. 11/11 "' •. ,Aolt. R ,,,.R,,- zz.nfHJ 0""', 2 ... "·S. I: ...(.':.(.' .... C ....- --.II I Il/.• 11/11 ... 'M,.rilu. S,- ,".,.1.1. I .. ;U" . R 1:--2(1(1 till"" -I . e . ;" ..o,u,1I ",,. (.'. - HI) Illl/.• #1111 ... ,.,.",,,. T.- - A .tlio 0.1,.1 I ••••. ",/(/ d"" , .. 6/16 R Ir--3fnfl 011... , 2 KO. '''1. (.'.-·10 Il/. . 3SI/II.• Iu. '8,.-i,0. dIN'. 311 ...1" (Dy •• A·llIl 0 •• q.iv.1 R ".R n-- 41n,Q()t} 0""'. �l KO. ,.,t. (.'.• C ••• C .. ,C •• -.lS Il/. . 11111 ... ,.,.. ilu. T :--1'0"" I ..... SSO·O-SSO p. (o! 1110 "' •. ; S p. R ,.-/fIlJII ""." 1 ... "".f. (.'. - 11/(/ Il/. . ISII ... du. r.,.ri,u. III 2 ,.Ht'I; '.3 P. � 3."" 1. 1t • .,....ISII d .... 1 ... 'rI. (.' ..- -11/0 1l1l/. . 11II1 v. r.,ori/Mr V •. V,-6SN11.b. R ....R -'�/fHl.(JOO 0""', " KO. ,." •• C,,--'OJlI.• -ISO 11'. ,.Irc. ('.,.dlu,. V"V\-J�U I ..b. .. 1 ft,,- 1011 d", , C,,- III /10/.. I.HI 11•• Iu. ,.,.,it ... V.' 1;7.31 1I::.II:.-·lIlfl oAHt. 1 ... rr •. (.',,-10 !*I.o 61/(/ ... du. ca�"'i/o. /for-S.,.,.I. . '.1 p ••• '."'" ., ••• IS .. , .. tI R -",Ru-IO(IO " . '", �/, •. " •• C •• -. SII Il/.• •�II ... • 1". '.' ....i/ •• tI.,., .. I., (A "" .. 11. 'NIliS)

June. 1961 10S 9 way to opUmize the high-frequency re· prodUCing that note In the ftrat place, • I1 sponse of the amplifier is to use an oscil­ One word of caution: the use of vt'1') Z7 loscope when making adjustments. In large bypassing values, 1000 flf, (m' ('X, o I1 �. practically all cases, bowever, thel'e will ample, may again lead to instability. IHOIIE be no audihle dilTel'ence between results �S Other Improvements tit I derived from u!!ing the given values and 04 I those elerived fmm inclividual trimming. A Williamson amplifiel' incorpcmll inl: If S The low-frequency transient response the modifications already ou t1incc I Idll IJ AfTEII 2 of thc Williamson is less critical and sounrlgood indeed but thel·c i!l one alllJi, 1./ e8!lier to !!traighten out. Many ampli· i I t onal thing you may do if you want to .v fiel'!!. tend to get "hloflpy'� when .hit hy be "elegant." Any "Ultm-Linear"-t�·Pe 0 0 2 4 _ • • 10 12 14 " " 20 U 24 ZII tiTgh--intensiiy. Iow-frequency sounds, Williamsoil- -even onc with a choke-in­ (r�":VE�;:_��� ) clue mainly to inadequate dceoupling of put. powel' supply -will display It dl'Op the "8 +." It is recommended that all of in "n +" dUI'inJ: extremely loucl pns.-­ FI,. 3. Improvement in performance of am ­ the stages he dccouplccl hy at least 40 sages. Thill hns a limiting elTeet on the plifier ulin, four 5"1·, in triode connection. ,.f., although in these amplificrs one of powel' output amI the amplifiel' may not the (sul'pl'isingly) critical points is the recovel' fl'om ovcrloads ,'Ill quickl,\- ;15 Amplifiel' No. 3 (sce Fi�. 41 uses two phase invel·tel' which should he mOl'e you'd like. The remccly illto dccoupl(' 1 he EL34/SCA7"s connected ns pentodes heavily decoupled. The now 3900-ohm output stn�e 11100'e heavily. It hns 11C'�n with a 250-ohm self-hias resistor and l'csistOl' decoupling the phase splittel· found that foUl' 9O-"f. elcetl'lllytics '·on­ 5000-ohm plate-tn-plate load. A�nin. no should he hypasscd with an elcctmlytic nceted across the center-tap of the out· special elTOI't WllS made to hlllnnce the whose value is in the neighborhoocl of RO put. h'anilfOl'mel' will pmvide supel'b drives. A 450-volt choke-input filter wns to 100 ,.f. Thesc chan�es will pl'o{luce a transient I'csponse. These capacitors used hel·e. stahle amplifier if you have uscd a go{){l hold enough of a chat'ge to supply the out pu t transfonnel' . output tubes with plenty of current ,lur­ Transient Response :-Jow that the nmplifiel' has heen sta­ ing short·eluration, high-intensity All of the fm'egoing will cut distortion hilized, let's try to furthel· impl'fI\·e t hc sounds and prevent n dmp in "B ·i-." The dramat ically, hut . we have yet nnother low-frequency t nlOsient I·cspnnsc. Thc amplifier shown in Fig. 2 does nnt ,lis· pl'oblem to face: tl'Hnsient I·esponse. Im­ capacitol' bypassing the cathodc resistor play the slightest change in "8 -I·" until provement in an amplifier's lI'ansient of the output stnge is usually sOl11e­ it is opcmting almost continuously in response is usuully more striking than whcre hetwecn 20 ,.f. and 250 ."f. Use of overload, Thc elTeet of this is most no­ impmvement in distortion. Now, the the 2;)()-,.f. (�apacitOl' hypasses the staJ:e ticeable on low-fl'equency transi('nls. high-frequency I·esponse of many \vil­ to at·ound 6 cycles, that is, the stagc is Also, it will give the amplifier practi­ Iinmson amplifiel's is ragged at hest and fiat to about 10 cycles. Low-pitched mu­ cally instantaneous overload l'cellvel·Y. it is nearly impossihle to give n set of sical waveshapcs contain nC1\l'-d.c. com­ Changing the bypassing of the out put­ values to smooth the high end of all ponents and in orelel· to hnnelle these stage cath{){le and incl'easing the out· these units. However, the values given ndequately you must have the output put-stage decoupling will likely pl1Kluce in Fig. 2 will be about right for an "UI­ stage bypassed to approximately 1/10th no audible changes unless you have n tra-Linear"-type \Villiamson using of the lowest r.·equency you wish to re­ really finc.spenkel' lIystem. \Vith such a either Stmu:or or D.I",a(:o lI'ansformel·s. produce. If you want to have go{){l tran­ system, the impl'Ovement is readily ap­ The RC values bypassing the plate load sicnt response helow 60 cycles, you have parent. of the first stage aren't too critical. The to bypass the output stage more heavily Of course, neither of the last 111"0 small capacitivc fccdback loop fmm thc than 250 /.f. Capacity of 500-600 /.f. m{){lificntions_ is o( much value if :hc plate of thc lower drivel' to the cathode (mnrle up of a couple of 300-,.f. elec­ phono preamp doe� not equalize ,·or­ of the first stal!c is also non-critical. trolytics in pamllell will hypass the rectly at the very low frequencies. t 'n­ However, the value of the capacitor out put. stage to ahout 3 cycles nnel will fortunately, many preamps don't. The should not exceed 100 JIIlf. Thc only give good transient response to nrounel boost applied below 100 cycles in the value that must be jugglcd with ex­ 30 cycles. It is n l'at·C speaker system RIAA curve is a help, but many record treme care is the capacitor acmss thc thnt can proeluce 30 cycles and it is n companiell do not always apply the pre­ feedback resistor. About 50 N'f. to 150 rare I'oom that will sustain that low n sumecl low-fl'equency pre-emphnsis. A p,p,f. is usual in this position. The only note even if the speaker is capahle of dillcussion of equalization nnel tonc-<:nn­ tl'lll is heyond the scope of this at'tielC'. �n Fig. 4. Performance of another amplifier modifi.d by author. This one utiliz.d a we will assume that you can get thc pair of El34/6CA7', with cathode bias and a choke-input filter in the power supply. lows into yOUl' amplifier. ! The modifications outlineel can Ioc 1 4 made for around $10 anel, in view of I he ! improvement in pel'fOl'mance, an� 11 ,," IS I1 j I worth the money, It is not the aiitholi··� I IJ 12 I intention to disparage the OI'iginaV"'i1- I i Iiamson nmplifier elesign-it has much I J I to recommend it, It is easy to build anll /BEFORE AFTER 10 7 by I·e-biasing the voltage-amplifiel� it Z 1/ becomell non-critical of tube selcetin n. � 9 Furthel'mOl'e, the "Ultra-Linenr" \·C'i'· or j I g 8 I sion is cnpahle of elelivering all the c1C';In '" , powel' nceded to drive any speakel' sy:<­ / , / o 7 tern except fOl' some of the less effid('1I1 I lE 6 I ---although excellent-bookshelf tYIII':<_ / 1/ The use of a chokc-input power supply IJ combined with the adrlitionnl hypassing 4 '/ and dec()upling will permit the "Ultl'l\­ ) V� Linear"-type Williamson to cleliver several watts mOl'C than ordinal'y cir­ V ./ V cuih'y permits. The distortion won't cx­ � V ceed 1�; until you l'each overload and I t.,....- --" will sound as though you have doublc<1 ...... - � the power output when what you havc o o 4'8 '0 12 14 " 18 20 22 24 21 28 SO S2 S4 36 S8 40 42 44 46 clone is improvc the overload pel'Corm· POWER-WATTS (E:OUIV. S INC: .. W&V[' ance. 106 ELECTRONICS WORLD