Home

showcasing a 20w push-pull amp cd to measure de

1. p 38 resides in Finland Ed Simon Book Reviwe Sound FX p 39 received his B S E E at Carnegie Mellon University He has installed over 500 sound systems at venues including Jacob s Field Cleveland Ohio MCI Center Washington D C Museum of Modern Art Restaurants New York The Coliseum Nashville Tenn The Forum Los Angeles Fisher Cats Stadium Manchester N H John Sunier Super Fidelity p 46 is a CD reviewer for Australian HiFi and Home Theatre Technology His website is Www audaud com 44 audioXpress 2 09 www audioxpress com cuit with a predetermined threshold R4 and C2 change _ This circuit is a compressor plain and toward ham radio enthusiasts where the The author weds an LED and a CdS extremely low by hi fi standards high The whole project seems more geared frequency turn down could make Morse code tones and voice more intelligible in _ the presence of a high noise background I m not familiar with ham radio conven is shown with input on the right and output on the left Overall the project is given the limited performance of the dont think that one of the applications would be music systems Cal Jonstone cjonstone1231 yahoo com J R Laughlin responds if you study the circuit a bit better to under stand how it works you will see that it does not compress or limit the peaks but simply as a volume control will do D1 and C6 store the approxim
2. surement preamp and the Fluke 189 true RMS meter are battery powered avoiding ground loops I measured the level of 60 120Hz line the measurement preamp output to an oscilloscope I used the scope s Tek tronix 475 inputs differentially channel 1 connected to the preamp output while channel 2 connected to the preamp s ground Then I used the scope s subtrac tion ability by inverting channel 2 and setting to Add mode The scope s ground was not connect ed to the test circuit except through the AC line plug s safety ground The 415V supply was also line grounded This prevented large common mode LMA HAS FILTER WATH tkhir 20khr BOUT NOISE Dud ve UNIT NHO iy Gain measured to refer noise to FET input LIT 2 Rs noise 0 4057nV Hz RMS subtracted out Fl Mi Ls _ dl te Laj ter EE e T a RRSRSERR LL 1 dl cli ft fey ee g SI e i LU D FIGURE 3 Noise input referred vs current 20 units 26 audioXpress 2 09 www audioxpress com multiplied by V2 for each FET s noise PE Ji Tiia ete a ETa which could occur if the 15V supply and or scope ground were floating NBW 20kHz 100kHz Noise with Q01 sample 7 out UV RMS innV RMS In average nV VHz 0 789 per FET 22 31 223 1 140 2dBV to 136 7dBV The average is 138 4dBV average nV VHz The observed AC line ripple 1kHz 100kHz 24 19 247 9 CONCLUSION was at leas
3. Noise each FET 1kHz 20kHz Avg mW AN Hz Noise 1kHz 20kHz nV RMS Noise ugr nV RMS 00 MI olol amp Go Ph 6 30 131 1 Vdd 10V Rd 509 802 RS 10 0002 gates grounded 10 877mA average 1 1646nV VHz average both FET sections paralleled combined current Table 2 Linear Systems LSK389B TDO71 Dual JFET Bias and Equivalent Input Noise Ship date 9 22 06 date code 0447 Noise Sample 20Hz 20kHz no V nV RMS 372 5 Lot JF 300 4 3 Test Date 12 4 07 Noise each FET 1kHz 20kHz Avg mW AN Hz Noise 1kHz 20kHz nV RMS Noise oN nV RMS 278 1 314 3 2 195 219 4 2 177 108 8 0 957 108 1 0 951 Tested 12 25 06 Vdd 10V RD 509 802 Rs 10 00Q gates grounded www audioXpress com both FET sections paralleled combined current a 3dB reduction But I also use a 10Q common source resistor for both DC bias stability and a return point for nega tive feedback from the AD797 op amp which is cascaded with the JFET The thermal noise of this 10Q resistor at 25 C is 0 4057nV VHz which is RMS added square root of sum of squares to the input referred parallel FET noise AS USED NOISE EXAMPLES I measured the FET s noise voltages in a circuit Fig 1 similar to that in the pre amps then calculated backwards RMS subtracting the 10Q resistor noise and then multiplying by V2
4. and with forms of automation Today this is the conclusion of the studio record ing process In the days of vinyl there would have been a chapter on actually getting the sound to fit into the groove It is assumed the digital files from the studio can be transferred unmodified to the distribution media with the current technology This book although clearly aimed at current pop music recording techniques offers many excellent explanations of basic to advanced audio theory and tech nique Much is applicable to live sound and the basics should help many audio philes better understand why they hear the things they do Case does not cover many of the other recording issues in this book There is very little on microphone technique The actual selection of the gear and its inter connection is reasonably avoided because that would limit the life of the book So in one sense the title is correct it is not intended to be a complete recording a Mustard Caps Mustard Laps guide It does contain more than enough _ information that this book with the use _ of the references could be the core of an Case then moves on to two special 3 excellent modern audio education _ The one shortfall that bothered me is the lack of attribution for some of the _ data presented As an example figure 3 1 isa very valuable chart of audio thresh should guide many of the readers
5. be more mumbles a bit over 8W Uh oh _ with the little heatsink I used in Example _ What to do A higher voltage trans former would work of course but Pm the chip will be sizzling at 168 C voltage regulator needs to see at its input already using a 12V AC transformer to get 6 3V DC higher seems extravagant _ Instead the best approach is to increase sink The Wakefield series I specified for the filter capacitance What if I make both big caps 4700UF sistance but with this kind of power I plus 100uF from the little electrolytics in parallel Now the voltage at the first regulators off board The LM317K and the next lowest value Putting that much current through a 1 0Q resistor will dissipate 1 21W so R8 needs to be a hefty one that s why I left room on the board for 5W dropping _ resistors Come to think of it I should probably juice up R5 too there s just enough room on the board for a 3W Pa EXAMPLE 3 THE PHANTOM WALKS For the third example I m going to use regulator s input will now be 0 17V AC 2 nasonic metal oxide resistor With these values the ripple at the NEW REPRINT FROM COLON The Art of Linear Electronics by John Linsley Hood a level I can live with What about heat At high line voltage the worst case regulator dissipation will 1 the temperature rise will be 128 C and if the ambient temperature is 40 C This clea
6. the mixer in the sun I have I design for multiple worst case scenarios and my NOT QUITE NOMINAL Ion the real world though not everything runs at nominal ratings Resistors have tolerances and so do regulators reference supplies I ve specified 1 resistors if R16 is 1 low then its real resistance will be 198Q If the regulator meanwhile is running at the high ends of its Vref and adjust terminal current tolerances then the rest is to light the metering LEDs Vref will be 1 3V and the reference cur Pll refer to part numbers for the supply 3 rent will be 5 Iref hi Vref hi R16 lo Iadj 0 00667A supplies very seldom fail or misbehave EXAMPLE 1 THE COMPRESSOR This was the ur circuit the one for which the board was conceived It s pretty stan dard stuff a dual channel compressor which requires a maximum of 0 25A from 15V supplies Seem like a lot The compressor itself pulls only 65mA the supply will use the same values The first step is to specify a few givens I m using two separate transformer wind ings so DU use D1 D8 to produce two bridge rectifiers R4 will be 1Q C4 will tentatively be 1000uF and C13 will be 28 audioXpress 2 09 3300uF If one cap is bigger I like to put it in the second stage of filtering to mini mize turn on current draw 2 An LM317 drops a nominal reference voltage of 1 25V
7. The AP2 CO 12dB octave is shown again shows a phase inversion occurs at the apart so when summed with normal tweet er polarity they cancel producing the in famous second order notch With inverted tweeter polarity the system sums to unity because now the LP and HP are in phase Two equal in phase signals sum to twice each signal or 6dB So all pass second or der COs cross over at 6dB down Note that the system phase shift shows a phase rever sal occurs at the frequency of the response polarity does the system show zero degrees phase shift and is thus phase linear in Fig 3 The LP and HP phases are 180 notch with normal tweeter polarity The B3 CO 18dB octave is shown in g Sai LR VU Bi 50 100 200 500 1 50 100 200 500 50 100 200 500 ep 100 200 500 A Frequency FIGURE 1 Performance of first order Butterworth crossover 42 audioXpress 2 09 Fig A The LP and HP are 270 apart which because phase repeats at 360 is the same as 90 apart Thus again the two sum to is clear Odd order COs LP and HP outputs 1 414 times each signal and must thus cross ower at 3dB Both tweeter polarities sum to unity Note however the inverted tweeter system has much less phase shift across the frequency band than the system using frequencies Review of the second order CO Finally the LR4 CO 24dB octave is shown in Fig 5 The LP and HP are 360 when the CO d
8. voltage than the level at 1kHz There fore the three noisiest of my ten units were significantly out of spec MY TWO APPLICATIONS THUS FAR I used this JFET in my Low Noise Measurement Preamp aX April 07 p 26 and in The LP797 Ultra Low Dis tortion Phono Preamp aX Sept 07 p 6 I was concerned particularly in the phono preamp about consistently ob taining the very low noise performance these JFETs are capable of Note that the ad is titled InV Low Noise Dual JFET 24 audioXpress 2 09 FAST FORWARD 14 MONTHS range median 1 1625 average 1 1646 On November 29 2007 I ordered 20 units of the LSK389B TO71 I re ceived them three days later lot no JF300 214 4 date code 0551 As Table 1 shows the 1 20kHz averaged noise density with I per FET ranging from 4 4 6 8mA more on this later was range 0 941 1 84nV VHz a 3 35dB The closeness of the median and aver _ age values is indicative of a well behaved symmetrical distribution more on this also later Table 2 shows the four units left from the 2006 group In both my measurement and phono preamps I use both FET sections in par allel this divides the noise voltage by V2 TABLE 1 Linear Systems LSK389B TO71 Dual JFET Bias and Equivalent Input Noise Ship date 11 23 07 Date Code 0551 Noise 20Hz 20kHz nV RMS Sample No Lot JF300 214 4 Test Date 12 4 07
9. Inf to 0 dB Insert new waveform at End of file w FIGURE 3 The Simple Synthesis module in Sony s Sound Forge will generate a sine wave at any frequency for any duration at any level ference relative to 20Hz Those results are plotted in the second column The maximum Red Book high frequency recorded at a level of 10dB which is REF TO 10dB The third column gives those levels relative to OdB in which case 20Hz is now at 10dB Finally the fourth col umn lists the track numbers for each test sets of tracks 1 through 28 are recorded flat at 10dB Tracks 29 through 56 duplicate the previous tones but with levels indicated in Table 1 I produced the test CD with Sony Creative Software s Sound Forge ver sion 9 0 the digital audio editor I use on an almost daily basis www sonycreativesoftware com Sound Forge has a Simple Synthesis function that al _ lows you to generate sine waves at any _ frequency for any duration you specify Fig 3 I used Simple Synthesis to gen State of the art engineering Extreme reliability Joyful souna Auricaps are precision wound and are considered by many to be the finest capacitors available Stereophile June 06 Peter Breuninger ovale flon simply the highest re capacitor ever made audience OEM amp dealer inquiries invited Call 800 565 4390 www audience av com audioXpress February 2009 33 er
10. Mary in Kevelaer Germany is a monumental instrument with 135 stops The opening five movement Sonata by Guilmant No 5 Op 80 is over a half hour length and is one of the composer s eight such organ sonatas It blends ancient and modern elements and ends with a grandiose final double fugue My pick of the disc is Bo llmann s Gothic Suite which also harks back to earlier music but avoids liturgical chant or even counterpoint After an Introduction and Chorale a minuet and a prayer to Notre Dame the work ends with an absolutely spectacular audiophile ecstatic Toccata finale which closes out the musical experience in stunning fashion AntonBruckner KZ Bruckner Symphony No 9 in E minor Suisse Romande Orchestra PentaTone Classics SACD PTC 5186 030 This incomplete it lacks a finale masterpiece was the shy Austrian composer s final symphonic utterance Ihe most modern of all his works the second movement Scherzo with its incessant hammering rhythm could fit right into the futurist mechanized music style which developed During Mozart s short lifetime there was a transition from the hegemony of the harpsichord and clavichord to the new pianoforte Christofori developed the gravecembalo col piano e forte while in Germany Hebestriet created his dulcimer inspired pianos known as pantaloons The vis a vis was a unique instrument with a harpsichord action at one end and a piano with bare wooden hammers at the other The Sonata
11. Solen split which is down 6dB at the crossover point It is supposed to reduce the problems of a first order filter Any help or comments would be greatly ap preciated G R Koonce responds Mr Mogavero has asked about the incon sistency in the crossover point between the first order 6dB octave and third order 18dB octave crossovers COs shown in my article suspect many readers may not be familiar with why the various CO orders have different crossover points so let me give a brief review will discuss the first order Butterworth B1 the second order all pass AP2 the third order Butterworth B3 and the fourth order Linkwitz Riley LR4 All of these COs Vincent Mogavero VMogavero aol com are capable of summing to unity across the frequency band the definition of an all pass CO In all pass configurations they do not all have the same system phase characteristic Fach CO is covered in the accompanying multi graph figures The plots are the elec trical outputs of the CO when loaded with _ fixed resistors You may alternately think of the plots as representing the on axis acous tic output of an ideal coaxial driver The top graph is the low pass LP mag nitude output solid line and the high pass HP magnitude output dashed curve The second graph shows the phase shift for the LP and HP outputs The third graph sho
12. T BOOK REVIEW STUDIO RECORDING TUTORIAL cre aaa Ae yr pe February 2009 US 7 00 Canada 10 00 Tube Solid State Loudspeaker Technology TUBE PRE DESIG te RAN O A M e Miva fg LE Length 0 008 tn 600 0 secorets Kb VV TM wa mm Pe CH amome 4000 0 SE E z EE GO D rma mes nnen es Ss tnt regan 3 01 02 Soe d SS O SE St DE Sg T SE Ss E Se S ge FF H a Seer S a NN sees ons E SE St K E SS S e www audioXpress com The ImPasse Preamplifier This author unveils his design of a tube preamp to drive the Pass F4 and other low to unity gain power amplifiers t all started with the Burning Amp Festival BAF 07 in San Fran cisco My friend was building a balanced version of Nelson Pass s First Watt F4 and asked for a recom mendation on how to drive it The Pass F4 is a unity gain power buffer meant to take a high level output from a pre amp and drive speakers with it an in teresting concept The schematic sug gests that the F4 can swing about 20V 40V peak in balanced mono A diagram of the balanced mono con amp which ought to be low distortion and low source impedance to reduce the criticality of interconnects So rather than recommend a preamp for the task I took the totally impracti cal step I don t own an
13. ct a a e e eee e e e i Df i J SES RAR EE Be HE FIGURE 7 Sony s CD Architect 5 2 allows full Red Book PQ encoding and editing in cluding the addition of a pre emphasis flag to any track Jan 09 issue p 32 was produced from The worst errors are at the top of the to 20kHz The NEW OLD COLONY SOUND LAB hal eo CATALOG 3 Ek lt eae hi ts rc OLD COLONY SOUND LAB You ll find Over 150 books and CDs on audio technology 3 wes New test CDs 7 NTIRE Software for design and measurement Sound Strobe and more test equipment ON LINE AT WWW audioXpress com Old Colony Sound Laboratory PO Box 876 Peterborough NH 03458 0876 USA Toll free 888 924 9465 Phone 603 924 9464 Fax 603 924 9467 E mail custserv audioXpress com FIND THE ENTIRE PRODUCT SELECTION ON LINE AT www audioXpress com www audioXpress com 36 audioXpress 2 09 www audioxpress com PRODUCT SELECTION data I measured with this test CD and spectrum where they are unlikely to _ shows excellent accuracy 0 09dB 20Hz be audible I also measured the DAC1 USB using the Hi Fi News and Record Review test disc mentioned previously The HFNRR disc showed 16kHz to be at 0 399dB left and 0 451dB right relative to 1kHz The measurements with my test CD showed 16kHz to be at 0 002dB left and 0 058dB right rela tive to 1kHz The nearly half a dB error shown by the HFNRR test
14. dB Ee 0 186 00 00 00 000 00 33 36 792 00 33 36 792 FIGURE 4 The FX Volume plug in supplied with Sound Forge allows level adjustments with resolution to 0 001dB umn When you finish if you zoom all ume functions that can boost or cut will look like Fig 5 one under the Process menu allows adjustments in 0 01dB increments the way out your Sound Forge screen 3 Ge You can then use Sound Forge s Sta tistics function under the Tools menu to check the level of each tone relative to OdB Fig 6 Highlight nearly all of at 1 65dB but don t highlight the silence on either CD AUTHORING With Sound Forge you first produce tracks by putting markers in the file where you want them You also need a marker at the end of the file When youre all done right click on the Re Table 1 Pre Emphasis Simulation Levels the tone in each track one at a time _ side Then click on Statistics and look 4 You ll need to highlight each tone from track 29 through track 56 and be the same and should also match the in the LEVEL REF TO 10dB col level given in column 3 of Table 1 for _ that particular tone at either Maximum sample value or Minimum sample value They should LEVEL FREQ REF TO 10dB 20Hz 0 50Hz 0 001 100Hz 0 004 500Hz 0 096 600Hz 0 138 700Hz 0 186 800Hz 0 241 900Hz 0 303 1kHz 0 37 2kHz 1 29 3kHz 2 43 4
15. sounds richer gutsier and less strident the low bass support is stronger and the 19 songs are in German but complete lyrics in English are in the note vital over structure of Bruckner s massive blocks of sound flows more naturally booklet Their version of the music from Rossini s Barber of Seville must and smoothly be heard to be believed The sextet is arrayed in front of the listener and sings a variety of classical lieder folk songs pop tunes and originals Art Pepper The Way It Was Contemporary Records Mobile Fidelity SACD UDSACD 2034 A lavish reissue on SACD from Mobile Fidelity is always a special thing and when the music attains such heights as it does on these early Art Pepper sides jazz lovers can t lose particularly since Pepper himself dictated the liner notes Some odd spatial placements occur as happened in the early days of stereo when engineers were feeling their way with it On the first tracks Pepper is playing cozy alto sax over on the left channel together with Warne Marsh on tenor while the drums and bass are way over on the right hand channel with absolutely nothing in between Then on some of the solo Pepper tracks he is heard coming from the phantom center channel Many tracks run more than six minutes leaving plenty of time for some great solo work The sound is clean the tunes are all classics and the individual approaches of the four different pianists make for added interest shakuhachi Neptune has bui
16. 0 en goyeng OC W N 20 DOH CO CEO Det a Ce DOS OCT Cp CEO Det D Ce Dr L t LSC Oise Dei DA Dream DO cL Dee a CSc Lin DO AAs Ori oF COLE 0e kiza 00k Lk L o biG GIL 0 MH Oe Gr er DOES OOF Grenn te Glee IOC Gb sie DM GeO DEO Dei oo OL bes OD e Cc eae DOA Oe M0 DITIT DOT ee Oe n ent rn _ according to the Red Book 50u5S 15u5 gions Sound Forge comes bundled with CD Architect 5 2 a professional disc at time constants wont tell your CD player the disc by your CD authoring program CD Architect allows you to add a pre the regions you ve already produced Open the WAV file with CD Architect s Open Media function and save the file FINISHED TEST CD as a CD Architect CDP project Click on Track List in the lower window emphasis flag to each track individually Once you ve finished adding the em phasis flag to tracks 29 through 56 save SBA Ts FREE TO UL EC ZEKE akut EFA IO eB AA SG F E SH SE E LA di ALIH Sbt Seen eG PO E FIGURE 5 The Sound Forge screen with the completed test CD Tracks 1 28 are re corded flat tracks 29 56 have been adjusted in level to correspond to the Red Book pre emphasis standard 1mV After making the voltage mea surements I convert them to dB in a to apply the correct de emphasis The Log E1 E2 pre emphasis flag must be recorded on vides full Red Book PQ encoding and spreadsheet using
17. 4 8 and 10 low cost full range drivers All exhib ited a remarkably similar series of im pedance peaks followed by a number of lower impedance ripples The bass driver used to assemble my ESP_replica network peaked with 843Q at 500kHz followed by multiple much smaller reso nances at 1 4MHz 2 5MHz 3 63MHz before finally settling to a reasonably high frequency For this article I mea sured impedance and phase angle of a representative range of speaker drivers from 1kHz to 10MHz The T27 tweeter used to assemble my ESP_replica circuit was peak reso nant with 825Q at 2 4MHz followed by a low of 105Q at 7MHz to 390Q at 10MHz A Kef B110B speaker reso steady 120Q A Goodmans 250W 18 power bass unit also exhibited multiple resonances with 495Q at 400kHz 138Q at 800kHz 400Q at 1 4MHz 272Q at 2 05MHz 360Q at 2 5MHz 213Q at 3 6MHz and finally 160Q at 6MHz before set tling down around 100Q at 10MHz While each of these plots differed all measured speaker impedances increased up to a first major resonant peak re sulting from the inductance and self capacitance of the voice coil and at fre quencies usually between 500kHz and 1MHz followed by a series of notably lesser impedance troughs and peaks like those found measuring most inductors By 10MHz all resonances had flattened _ like the measured values resonant peaks to
18. 826 5432 Yard Sale For Sale 2 EICO 1030 power supply with NOS Mullard GZ33 all NOS tubes and extra NOS 6L6GC tubes 50 each plus shipping ashby123 comcast net High End Driver Sale for the DIY Speaker Builders Large selection of high end high quality woofers midranges and tweeters from Acuton Audax Eton Focal Morel Parts Express ScanSpeak SEAS Usher and Vifa All half price All new A few used briefly in loudspeaker system development Thiele Small parameters and suggested crossover circuits available Also available a number of high quality cabinets to make complete speaker systems Laboratory grade microphones and measurement systems ACO Pacific 1 2 free field microphone system including the 7012 12 free field mike 4012 preamp and PS9200 power supply Recent calibration file provided 975 Neutrik HF882 mike with calibration file 250 Also CLIO and DAAS PC based electroacoustic measurement systems Write Joe D Appolito 34 Rust Pond Rd Wolfeboro NH 03894 or e mail audioltd metrocast net for details Yard Sale is published in each issue of aX For guidelines on how subscribers can publish their free ad see our website audioXpress February 2009 37 20W Push Pull Amp ere are some photos of an amp have built from your magazine article A 20W 260 Amplifier Glass Audio 5 00 by Joseph Still made a few modifications choke filter in the power supply better compone
19. 9465 e mail custserv audioXpress com ai nNN S YPaOe03S030 0NN ACOustics Begins With ACO audioXpress February 2009 23 TTT EIIE Ey Dennis Colin Noise Measurements of the LSK389B Dual JFET Good noise performance with reasonable consistency was observed on a shipment of 20 units he LSK389 dual JFET by Lin ear Systems 800 359 4023 or 510 490 9160 Fax 510 353 0261 www linearsystems com has been advertised for some time in audioXpress See ad in this issue on p 41 In this ad the voltage noise density input reference at a drain current I4 of 3mA is graphed at 1 3nV VHz at 1kHz and 1 0nV VHz at 20kHz I ve heard some concern about the consistency of this part On Sept 22 2006 I ordered ten units of the LSK389B TO71 The B category is an is saturation drain current at zero gate source voltage range of 6 12mA There s also the A with 2 6 6 5mA and the C with 10 20mA Of the ten units three had a volt age noise density from now on called simply noise ranging between 2 and 3nV VHz while the other seven were between 1 0 and 1 8nV VHz This was a measurement averaged over 1kHz 10kHz Meanwhile the datasheet specifies 0 9 typical 1 9 maximum nV VHz at 1kHz From the downslop ing noise versus frequency graph shown in Linear Systems ad the 1 10kHz av eraging I used will show a lower noise
20. CD is suspicious In the _ datasheet for the AD1853 the DAC chip used in the Benchmark DAC1 USB Analog Devices specifies the de emphasis error as 0 1dB The measurements made with my test CD show the Benchmark DAC to be well within Analog Devices tolerance After sending Chuck Hansen a copy of my test CD he said he would use it for his de emphasis measurements I thank him for his feedback and words of encouragement during the prepara tion of the test CD and this article I hope other readers will find this disc as useful as we have _ The De Emphasis Test CD can be pur chased from the author for 20 each including Media Mail shipping in the _ US Send a check or money order pay able to Gary Galo 211 May Road Potsdam NY 13676 aX H Classified VENDORS High Performance kits Audiophile components Custom designs Custom Assembly www borbelyaudio com In North America LBAudio Les Bordelon Ibordelon roadrunner com In Taiwan TS audiolab Tai Shen Lee tsaudio lab msa hinet net AudioClassics com Buys Sells Trades Repairs Appraises Mcintosh amp other High End and Vintage Audio Equipment 800 321 2834 Vista Audio Radii Audio Limits Trafomatic Tube amplifiers kits custom transformers www engineeringvista com Class A Chassis New ezPower Chassis DIY power amplifier chassis includes heatsinks www designbuildlisten com Read about the ODDWATT 225 class A tube power amp at
21. DS WE ee eee eee r 1 O a 0 100 200 500 1k 500 1k 100 200 500 1k 10k 20k Frequency FIGURE 4 Performance of third order Butterworth crossover the crossover points are not what ideal CO theory predicts know nothing about the Solen split CO and can therefore not comment on it hope this answers Mr Mogavero s ques tions and clarifies the different CO points wish Mr Mogavero good luck with his speaker project a 50 100 200 Dk 10k 20k i ki LH i 20 50 100 200 500 tk 2k 5k 10k 20k en 100 200 500 ik 2k 5k 10k 20k Frequency FIGURE 5 Performance of fourth order Linkwitz Riley crossover GUITAR AMP to produce the desired system response I have some questions for Mr Joseph _N Still about his tube integrated ste reo amp Sept 08 I am a musician and would like to build this amplifier to use with my tube preamp Here are my questions 1 Can the bias system which is cath the correct feedback is 9dB I m very sorry ode bias be replaced by the system you used in your 60W ultralinear amplifier from audioXpress 6 04 on a single chassis which is available from Mouser Electronics 800 346 6873 The size 2 Can I bridge both output trans formers with a switch to get a mono output parallel them when not TUBE AMP _ This is in reference to Alexander Arion s I have all the parts to build
22. F4 of designing and building one for his demonstration at BAF 07 And of course it needed to be called The ImPasse THE ROAD NOT TAKEN I will first examine the design require ments In the balanced mono arrange ment your preamp must be able to swing 20V peak on each of its output polarities and do that driven from a records that are cut at a lower than LEFT CHANNEL BALANCED OUTPUT PREAMP 3 very low even with relatively large gate typical line level source 2V RMS or 2 8V peak To accommodate CDs and 2 resistors This relieves the preamp from the need to deliver much signal current it just needs to swing many volts very normal level you should build in some extra gain arbitrarily and conveniently arriving at a target gain of 26dB A 201 In the balanced mono connection the power amplifier input impedances are identical and symmetrical Thanks to the FET input their values are also peak into an 8Q load single ended but cleanly For me at least the requirement to swing a lot of volts at low current with tubes And a simple circuit suggests _ what s even more compelling is that tube stages that give balanced outputs at high voltage levels are already well known somewhat arbitrary the intrinsically _ low gate current of a FET means that voltage offsets due to gate current can be nection for the F4 is shown in Fig 1 A p
23. a relatively consistent moderate im simulation You may wonder why I chose to use pedance and phase angle by frequency into the Z_block as shown is quick er simpler and most important error the Z_block to represent my test induc tors Why not simply model their sche _ PHANTOM Powered Measurement Mic system 7052 Type 1 5 Titanium Diaphragm 3 Hz to gt 20 kHz lt 20 dBA gt 140 dBSPL MK224 Optional 4048 Preamp 2 uV A gt 5 Vrms 4 Hz gt 100 kHz 18 to 60 Vdc Fr 30 0 00 9 E sa wi O a sl Superior F Long term Stability IEC1094 Type Temp and Humidity Performance I Now in Stock 7052PH MK224PH ACO Pacific Inc 2604 Read Ave Belmont CA 94002 USA Tel 650 595 8588 Fax 650 591 2891 www acopacific com sales acopacific com increasing with frequency quite un and troughs so this can lead to false conclusions However you must always remember that large value capacitors are series reso nant at audio frequency The Elna 4700u 63V aluminum electrolytic used in my 7 5kHz Accurately measuring an inductor ca simply inputting measured values of im free producing the most accurate simu lations possible for any component or even a complex speaker connected via its Cyril Batemans extensive work on capacitors as originally published in Electronics World is now available on CD from Old Colony Sound Lab PO Box 876 Peterborough NH 03458 888 924
24. be applied prior to A to D conversion in order to overcome the limitations of the converters A complementary de emphasis equalization curve was applied 1980s either for their own label the en 30 00 dB 25 00 dB 20 00 dB phony Mahler cycle for example or _ discs they made for other labels Music and Arts Programs of America had mam of their early CDs manufactured in Japan by Denon In recent years many manufacturers of CD players and implementing playback de emphasis Oct 2007 aX is a case in point This is a problem for those of us who have been collecting CDs since the 1980s I believe that all CD playback hardware should be backwards compatible entire process had become a moot point due to improved linearity of both A D RED BOOK STANDARD and D A converters Very few CD manufacturers actu specified as 50uS and 15uS corre sponding to frequencies of 3183Hz and CDs I own with pre emphasis are discs manufactured in Japan by Denon in the _ The Red Book pre emphasis curve is shown in Fig 1 Time constants are 10610Hz Relative to the low end of the spectrum the 3dB point for the boost is 3183Hz with the boost shelving at tire Eliahu Inbal Frankfurt Radio Sym 10610Hz 15 00 dB 10 00 dB 5 000 dB 0 000 dB 5 000 dB 10 00 dB p p KN 3 15 00 dB 20 00 dB 10 00 Hz 100 0 Hz 1 000kHz 10 00kHz 100 0kHz FIGURE 1 The Red
25. between its output and adjust terminals If R16 200Q then it will draw a total nominal reference cur rent of L Iref Vref R16 Iadj 1 25 200 16V This won t hurt the audio circuits What s Iadj That s current coming out of the adjust terminal as specified in the opposite direction Vref is at 0 00005 0 0063A 6 3mA in the datasheet For an output voltage of 15V the ad limit of 202Q and R16 is at its low limit of 2188Q There s no minimum spec for adjust terminal current so Pll use just terminal should be at Vout Vref 15 1 25 13 75V That voltage comes from dropping the reference current through R13 which 8 Iref lo 5 99mA should be R13 Vadj Iref 13 75 0063 2183Q i The nearest E96 value is 2 21k If dhe 2 current is 6 3mA the power dissipated 2 in R13 will be O88 WW A quarter watt metal film resistor should be able to handle that without problems So far so good 1 3 198 0 0001 6 67mA If meanwhile R13 is 1 high its real resistance will be 2232Q Pass the above current through it and you get www audioxpress com 0063 2210 6 Vreg hi Vref hi Iref hi R16 hi 1 3 00667 2232 16 2V And R13 will dissipate in the worst case 7 P Iref hi R13hi 0992W Still pretty reasonable for a quarter watt resistor So under the worst combined conditions the
26. current and drop a lot of volts from the B rail you ll see later why the B rail needs to be fairly high about 350V So because the CCS has such a high impedance adding some series resistance doesnt much affect its operation but R4 will dissipate some of the power necessary in the plate circuit This will eliminate the heatsinking requirement for Q2 which would otherwise need to remove more than a watt of power I chose its value to drop about half of the necessary voltage It serves a more subtle purpose too it isolates any of the stray capacitances of the power FETs from the plate circuit For thermal rea sons it should be generously rated a 2 or 3W unit is ideal Volume controls are a common weak point in analog pre amps You should use a high quality one a stepped attenu ator would be my first choice For my prototype I selected parts on hand wherever possible so I used an Alps Black Beauty Personal choice should rule here A too large value will interact with the Miller capacitance of the 65N7 to cause treble losses and too small can cause loading issues for the input transformer You need to pick the input circuitry as a team You must first and foremost cater to the whims of the prima donna the input transformer It is more than a bit fussy about how its secondary is loaded otherwise it will make you sorry you ever bought it And here s where you ll need to do a bit of experimenting if you use a different input tra
27. due to reduced Miller effect If you use a high mu triode such as a D3a as an input tube the input capacitance for a single ended source will be nearly supply rejection is excellent To get to the target gain of 20 you need a tube with a gain somewhat larger than 40 from a single ended source gain to each plate of a differential ampli fier is half the gain that the same circuit would have in common cathode Re TIS A GIFT TO BE SIMPLE _ Turning to a different topological option _well trod by power amplifier design or 12AX7 ECC83 Being more exotic voltage amplifier feeding a split load of the volume control assuming a low stricting yourself to easily available tubes you could look at 12AT7 ECC81 6SL7 you could check some of the high gm European pentodes triode connected in order to achieve the requisite gain with a low plate resistance Consider the common tubes Of the has the lowest plate resistance at 10 15k for normal operating currents the preamp s output impedance will be 20 30k which is somewhat higher than is FIGURE 2 Classic differential amp 6 audioXpress 2 09 _ seems that you ll need to follow the dif ferential amplifiers with a pair of buffers probably cathode followers so a well You can do somewhat better with the about 4k5 considerably better but still quite high But with care and attention some expense D3a is not cheap and
28. exotics and input capacitance is significantly lower Among common medium mu tubes it is the most linear in voltage amplifier service and the mu of 20 is spot on for this application FINAL ANSWER So given the design choices here I ve gone conventional and will use a 6SN7 as a single ended common cathode volt age amplifier and an ECC88 as a split load phase splitter buffer It s cheap easy ond order Not bad for a single device current through the LEDs will also be constant because plate current equals cathode current so any nonlinearity in and as good better performance than the exotics in differential mode And a whole two channel preamp can be done with two tubes The fact that I had a single CV1988 a rare British military _ version of the 6SN7 begging to be used _ did not influence my choice of tube here _ sive so an interesting and very viable no sir UU now move to the step by step practicalities of the design The input stage provides all of the voltage ampli fication and is hence the most critical _ pedance low noise and reasonably high regarding distortion and available volt age swing It needs to swing slightly red LED will typically show 1 7V drop with an AC impedance of 5Q or less at ly loaded a split load inverter has the same low input capacitance as a cath ode follower so can have a high input 2 for the 6SN7 you see that f
29. higher than the 1 165nV VHz figure averaged over Che 1kHz 20kHz band This reflects the ples 2 and 16 unit 2 has the highest the two paralleled FETs with the LOL sity of 2 278nV VHz averaged over the Then RMS normal low frequency noise increase of semiconductors FET and bipolar However as the A weighting curve _ shows the ear is relatively insensitive to Parts List 2 low level low frequency sounds A WEIGHTED NOISE This data in Table 1 ranges from 97 4 145 8nV RMS with an average of the 20 samples of 120 47nV RMS This is for the as tested configuration of two paral leled FETs and the 10Q source resistor Calculation of the A noise for a single FET without source resistor is difficult weight 2 OAD Frequency response 3dB BW 1 4Hz 800kHz of my 2007 order were comfort ably within the noise spec Linear Systems offers low noise screening for a fee I don t yet know the cost but if these 20 units are representative that shouldn t be necessary If you build the LP797 phono preamp I recommend using the specified DIP sockets for the JFETs Then by shorting the preamp inputs and measuring the output noise you can select JFETs for lowest noise For small quantity orders from Linear Systems it probably costs less to buy a few extra JFETs so you can select the lowest noise units than to pay Linear Systems fo
30. in B major is heard on the vis a vis An unfretted double strung clavichord based on a 1772 instrument is heard in three short Mozart selections Two very different square pianos also feature on this CD because Mozart especially liked this type of instrument The enhanced resolution of SACD makes it easier to distinguish the subtle differences ajoute Lo trompectte d BRASS QUINTET amp ORGAN Frommermann Music of the Comedian Harmonists Channel Classics SACD CCS SA 26807 The unique male vocal sextet known as the Comedian Harmonists was founded in Berlin in 1927 by Harry Frommermann That sextet combined crack vocal technique and arrangements which led the way to many later innovations in jazz and classical in the 1920s The last movement he actually vocal ensembles with a wonderful completed the Adagio is the longest of the three and abounds in many sense of humor and a charming presentational style The Dutch vocal intense orchestral climaxes It is so monumental sounding that it brings the group Frommermann named itself after the founder of the Comedian symphony to such a very logical conclusion that you dont miss the missing Harmonists finale at all This recording is superior to the Vienna Philharmonic SACD that Established in 2004 Frommermann has the same voice blend as well was one of my Best of the Year picks last year The Suisse Romande orchestra three tenors two baritones and a basso plus their pianist Several of the
31. one cially at low voltages limiting the source impedance at high frequencies Worse yet at lower drain source voltages the capacitance is very nonlinear 2 6SN7 FIGURE 4 Out line of the voltage gain stage www audioxpress com Ideally you would hold V constant And that is the function of Q2 which is connected as a source follower with the _ gate driven by the source of Q1 R7 is a gate stopper as before with no signifi cant voltage drop across it Q2 s source follows the signal so that the voltage between Q1 s drain and source is held constant Thus the drain source capaci tance is not modulated and the constant Considering those flat I versus V Vac across Q1 means that current is held more constant equivalently you can say that the output impedance is higher largely independent of the drain source voltage If the current tended to rise the voltage across R5 would rise which would tend to drive the FET toward lower current The opposite argument can be made for the FET tending to _ ward lower current So because of that Feedback effect I q is constant The complete schematic is shown in Fig 6 You can now start assigning com ponent values The gate stoppers value p 8 Pp is largely non critical a few hundred ohms will do Because I had a bag of 200R carbon resistors that was my cho sen value If you use 330R or 270R or even
32. regulator will be putting out a bit over any but it ll affect the design process What happens when everything goes its low limit of 1 2V R16 is at its high the typical figure Then sparing the math 9 Vreg lo 14 3V Pll use those numbers later OK I ve presented the worst case for regulator current draw it ll be 6 67mA Since the specified load cur rent is 250mA total current drawn from the supply will be 256 67mA Those are more significant figures than necessary let s call it 257mA or 0 257A Next you need to figure how much voltage the regulator needs across it to keep regulating properly known as the dropout voltage That comes from a chart in the spec sheet and interpolating a little at this current draw it wants at least 1 8V at a chilly temperature of 10 C I ve worked gigs like that too If the dropout voltage is 1 8V then you ll need at least 10 Vunreg Vreg hi Vdropout 16 2 1 8 18 0V to maintain regula tion under worst case conditions I normally add a fudge factor of 0 25V to allow for supply ripple so I really want 11 Vunreg 18 25V That s the bottom line number the raw supply must feed the regulator at least 18 25V COLD AND RAW remember that I m using separate wind Nominally 18V AC that is Unfortunately this is another spot US fluctuations of 10 in the volts coming out of the wall are pretty typical and in most of the c
33. style This chapter _ makes very clear to the journeyman the _ difference between older styles of record ing live recording and current studio practice what the student needs to add to his her knowledge to do useful work The first section of the book includes mathematically and explains why these principles concern us musically The au single explanation I have ever seen of this material Alexander U Case Sou mca Unlocking the Creative Potential of Recording Studio Effects f Focal The next three chapters cover equaliza _ tion compression and limiting then ex pansion and gating Although presented from a recordist s point of view these are a quite complete and understandable _ explanation of the technologies the actual devices the limitations of these processes and how to use them for both recording and to a lesser extent for live work Spe cific examples given of where and how to a much more hands on feel to the book Chapter eight on volume controls is probably the single most important chap ter in the book You would have expected The next section deals with time based effects The three chapters cover delay pitch shift and reverb A musical theory approach is taken to many of the expla nations of how to use these processes _ There are some charts relating tempo to time in chapter nine that will undoubt _ edly be a constant reference for some o
34. sym Frequencies Q feo araa a tco metrical WTW array assuming ideal TwoWay 2k 0 5 4k 0 5 1k Three Way 500 3k 0 5 1k 0 5 250 6k 0 5 1 5k drivers for a design CO frequency of 2kHz The magnitude anomalies are about 2dB and the phase holds to about 7 The approach can be extended to a three way by designing the bandpass to the same O and overlap as for the other sections Figure 2 shows the performance for a three way five driver WMTMW symmetrical ideal array The performance is not as good if the two crossover frequencies become close together This is a simple approach with low parts count that ultimately gives 12dB octave slopes but still wants drivers with good response overlap recommend using symmetrical arrays to limit lobing problems Table 1 shows the design pa rameters for the crossovers in the figures to clarify the approach hope all those working on time disper sion effects and limiting designs will publish so we may all make progress in this area Right now admit I m still confused about what constitutes the tolerable limit and how it varies with frequency and other variables The linear phase speaker article by Mr Koonce was of great interest to me be cause I am building an almost identical system I noted that the 18dB per octave crossover is 6dB at the crossover point while the 6dB crossover is 3dB Is this correct Also there is a 6dB per octave
35. the current to run two channels of pre amp will be 80mA or more ers consider a simple common cathode inverter Fig 3 You can couple the two stages in various ways direct coupling BC coupling or a combination of the two are all simple and effective three candidates the 12AT7 ECC81 verter has symmetrical source impedance from both the plate and cathode sides as long as the load is also symmetrical That source impedance is lower than comfortable Any cable capacitance will be liable to cause HF losses The other tubes produce a much worse problem It Contrary to myth the split load in that of a comparable diff amp being ap proximately 1 gm similar to a simple cathode follower For a common tube such as ECC88 for example Z be approximately 100 200Q a much more attractive figure than even the ex designed preamp based on that topology will have four tubes at a minimum i will otic differential amplifier Better yet the overall gain of the combination of grounded cathode volt exotics A D3a triode connected will have a plate resistance of slightly greater than 2k at 20mA operating current This will result in an output impedance of use for the first voltage amplifier stage ro include medium mu triodes such age amplifier and split load inverter will have double the gain of the same voltage amplifier tube in differential mode This extends the range of tub
36. the formula dB 20 De emphasis errors are similar to RIAA equalization errors in phono preamps Because errors are sometimes spread across an octave or more even errors of a few tenths of a dB can be au dible if they occur between 1 and 5kHz _ where the ear is especially sensitive The _ your changes in the CDP file and burn the CD You ll now have a de emphasis test CD that is far more useful than any _ published test discs that I ve seen To check your CD player or outboard DAC first play tracks 1 through 28 monitor _ ing the player or DAC output on an AC voltmeter with a dB scale These tracks should show a flat frequency response from 20Hz to 20kHz Now do the same _ with tracks 29 through 56 _ If your player or DAC supports the Red Book de emphasis specification you should also get the same flat frequency response you got with tracks 1 through _ 28 If you get a response that rises with frequency with 20kHz at about 9 5dB relative to 20Hz your player or DAC doesn t support de emphasis For the most accurate measurements I use a gt _ digital AC voltmeter with resolution to de emphasis graph Chuck Hansen pre pared for the Benchmark DAC1 USB review Fig 2 in his review published in MS Statistics Pre Emphasis Tes CD Vi wav FIGURE 6 The Sound Forge Statistics module allows verification of the level of each tone relative to OdB Use the
37. to obtain the per FET noise Two in parallel will halve the equivalent noise resistance The column in Table 1 labeled Noise 1kHz 20kHz shows the directly mea sured integrated noise voltage input re ferred over this bandwidth I followed the test circuit of Fig 1 overall gain stabilized at close to 100 by the feed back with the low noise measurement 3 preamp its gain set to 1000 The latter s selectable noise bandwidths were used to measure 20Hz 20kHz A weighted and 1kHz 20kHz noise levels Figure 2 shows the overall noise test setup Taking the 1kHz 20kHz values and dividing by V19 000Hz which is 137 84VHz gives the 1kHz 20kHz averaged noise density values for the configuration of both FETs paralleled and the 10Q source resistor In the 20 unit sample this ranges from 0 779 to 1 059nV VHz The equivalent thermal noise resistance 25 C ranges from 36 9 to 68 1Q Note that these resis tances would be 10Q lower if the FET sources were directly grounded JFET CURRENT IN TEST CIRCUIT Note that there s no overall negative feedback NFB at DC because of the coupling cap C2 But the 10Q source re sistor R4 provides source degeneration analogous to emitter degeneration in a _ bipolar transistor This simply means negative feedback and here the NFB extends to DC because an increase in FET current generates a greater voltage drop
38. www tdl tech com preamps htm Magnificent sound TDL Technology Inc 575 382 3173 ADVERTISER ACO Pacific Inc Antique Radio Classified Audio Amateur Corp audioXpress Subscription Old Colony Sound Lab Catalog Art of Linear Electronics Music CD s Testing Loudspeakers Audience Audio Transformers Avel Lindberg Electra Print Audio Co ETI Eichmann Technologies Hammond Manufacturing Flypex Electronics B V ccccissccsscuvacsecaesnes 43 KAB Electro Acoustics Liberty Instruments Inc 11 Linear Integrated Systems SUBSCRIBE AND SAVE Super Fidelity Music Review Column Debuts _ audio press Versatile Filter For omg Speaker Analysis IMPROVED HYBRID AMP DRIVER CIRCUIT MEET SYSTEMS DESIGNER EDUARDO DE LIMA a SEARCH OF TREASURES A SIMPLE WAY TO FIGURE TUBE DISTORTION SAVE NEARLY 50 OFF THE COVER PRICE Call 888 924 9465 or subscribe on line at www audioXpress com Madisound Loudspeakers Marchand Electronics Inc Mouser Electronics Mundorf EB GmbH Parts Connexion Parts Express Int l Inc PCBCart com Sencore SOOM PING wee ease eee R 19 Tang Band Industries Cold 7 Test Equipment Depot Tube Amp Doctor GmbH CLASSFIEDS All Electronics Audio Classics Ltd Borbely Audio Design Build Listen Ltd ENG Mistaiin EE 37 TDL Technology Kant Electronic and Electro mechanical Devices Parts and Supplies Many unique items www allelectronics com Free 96 page catalog 1 800
39. your Greek Triad article in the October 08 60W unit but I see you have a new 50W one that uses fewer parts I know this is a hi fi amplifier but would love ro use it in my guitar rig it would give me a clean warm tube output power amplifier This system has an adjustable bal ance bias system using them in stereo mosfetman20 yahoo com Joseph Norwood Still responds 1 Yes you can use the bias system from the amplifier in the 6 04 article _ 2 You can bridge the 8Q outputs by _ using a switch if the guitar speaker is 4Q If you use an 8Q speaker parallel the output transformer with 16Q outputs found a matched pair of 6550s in a fixed bias configuration gave a balanced output so you might choose to try this configuration before going with the more complex bias system The amplifier in the 9 08 article contains an error The feedback of 12dB is incorrect for the mistake assume you intend to build the amplifier of the aluminum chassis is 17 x 10 x 3 issue First I can t see the advantage of the cascode stage Maybe I just don t understand it well enough Second the coupling capacitor to the grid of the output stage is so big that the response would be down 3dB at about 0 6Hz I pity the output transformer Dusk Ball _ without any coupling capacitor so tran _ sients from switching inputs are applied directly to the grid Looks like a po
40. 0HG Kattegat 8 9723 JP Groningen The Netherlands 31 50 526 49 93 sales hypex nl www hypex nl audioXpress February 2009 43 Alex Arion responds Thank you for your inquiries about the Greek Triad The first stage is not a cas code but a SRPP which is good enough nor CDS cell with epoxy and shrink sleeve Silonex www silonex com makes a series of Audiohm optocoupler de vices that perform this same function tions perhaps that is why the schematic for the job selected the 6SL 7 6H9C instead of the classic 6SN7 because of 3 its M factor bigger than 6SN7 6H8C The coupling capacitor 1MF was se lected especially to transfer a lot of low frequencies bypassed by a 4 NF The output transformer is doing well without References With the part values as shown a steep attenuation of high frequencies through AD820 op amps and that LM386 I just out the circuit starts at the input where even at full volume control setting R18 and C9 reduce the HF 3dB point to just 3 8kHz It will be reduced even fur ther as the volume control is lowered _ By the time the audio signal makes it to the output the overall HF 3dB point is only 2 7kHz 2 Op amps U1 and U2 have limited gain bandwidth and should have small _ divides down the entire audio signal such tors R14 and R2 to avoid oscillation Any variations in R4 will change the LF 2 3dB po
41. 0Q resistor are AC coupled to ground the current being highly stabilized with a tered 15V supply The low noise mea attempt to increase the current This is useful in stabilizing the bias current I against unit to unit I vari ations The LSK389B is specified with an Ij range of 6 12mA a 2 1 range I didnt measure I on these 20 units but 3 I did on the previous shipment of ten units I was in spec on all ten units In this test circuit the FET cur rent ranged from 8 85mA unit 2 to 13 55mA unit 16 This is for both FET _ FURTHER NOTES ON TEST CIRCUIT The FET s transconductance is speci fied at 20ms typical at 3mA each FET for the two in parallel that would be 40ms With this value the circuit s over very little decrease in noise Also with a all feedback loop gain would be 12 4 21 9dB With this loop gain a 2 1 variation in FET sample transconduc tance say from 14 to 28ms per FET produces a 5 4 overall closed loop gain 0 46dB variation I observed much less variation The circuit has a 3dB band 2 width of 1 4Hz 800kHz Therefore the voltages from appearing at the scope sections paralleled so if the two FETs in the package were matched the range of per FET currents would be 4 425 6 775mA I had previously found that in creasing the current above 3mA produced minimum of 10V at the top of the drain resis
42. 112 Riii To PS GND R103 R104 TY R10 om C102 103 PS GND To GND www audioxpress com able current capacity and high transcon ductance Now how do you set up the output stage In a split load inverter the cur rent through cathode and plate resis tors is equal so the voltage drops across them are equal I want to run the cur rent high enough to keep the tube s transconductance high but not so much as to threaten reliability As with the first stage 8mA is a good compromise Now we need to choose the load resistors Too low and the distortion goes up too high and it s difficult to impossible to get 8mA of current at any reasonable supply voltage Because I ve used ECC88s in line amps with a 30k load and got reasonably From 350 D103 IC1 5 350 R107 R109 R110 R108 C104 To GND good distortion performance and 8mA its plate so direct coupling becomes dif lose 40V or so of DC There are several gives a 240V drop plenty to supply our ficult In order to get the output stage ways to do this for the sake of simplic swing requirements I will declare that grid to the right voltage you need to ity I used a voltage divider R8 and R9 the Official Value Splitting the load means that the cathode and plate resis tors R11 and R12 are each 15k with a drop of 120V or so across them at the nominal 8mA You can now see why the B needs to be at 350V or more R11 and R12 ea
43. 3 509 8 R4 9 99 R5 R8 180 R6 1 067K R7 46 5K R9 217 3 R10 549 4 R1 R12 10K s1 Switch SPST U1 AD797 with Q1 sample 7 1kHz gain 10 0032 100 027 regarding Q1 gates v 10 64V 20 1mA V 10 29V 8 6mA 01 Vd 4 748V io 11 56mA without knowing the resistor s ed noise and will not be estimated here _ But because the as tested configura tion has proven to have reliable low noise performance the above A weighted in put referred noise voltage measurements _ are relevant These levels range from 1kHz maximum out 5 14V RMS Vo DC 2 0mV Both FETs combined lot JF300 4 3 date code 0447 SUMMARY FROM TABLE 1 DATA Here I ll use the average values of the 20 samples Integrated noise voltages of bered with the LSK389B test samples source resistor are 145 355nV 20Hz 20kHz BW and 126 655nV 1kHz 20kHz BW RMS subtracting these gives 71 320nV this is the integrated total noise in a 20Hz 1kHz band Di viding this by V980Hz gives a noise den is an excellent product very suitable for low noise audio applications subtracting the 0 4057nV VHz thermal noise of the 10Q resistor and multiplying the result be V2 for each FET of the paralleled pair gives 3 170nV VHz average noise voltage density in a 20Hz 1kHz band for a single FET at an average current current correlation but obviously other of 5 44mA This is 8 7dB
44. 470R the operation of the CCS will not be affected it is important to have the body of the resistor as close to the gate as possible The value of R5 is more critical that s what sets the CC is current And the CCS sets the operating current of the voltage amplifier tube FETs vary A lot It is not unusual Those parts were all relatively similar with two or three outliers from the batch of 100 tested But I cannot guarantee that other lots of these parts will be so closely controlled So although I give the value of R5 that worked for this particular lot you will need to ad just its value for the devices you have to get the desired operating current To test the operating current and check whether the initial value of R5 must be raised or lowered you can use the circuit in Fig 7 This jig is powered by two standard 9V batteries in series of course you can use any other con venient power supply The current is calculated from the measured voltage across R divided by R ests Value A convenient value is 1k which will give an 8V drop at the nominal operating current Now the absolute value of the current isnt extremely critical but you will need to match both channels Although 8mA is the design current here going a little bit higher or lower say be tween 7 5 and 9mA wont ruin the performance There s one more little trick up my sleeve R4 The CCS needs to carry all of the 6SN7 s plate
45. 47R 0 5W KA WK o eee renee re ee 1M 0 125W Malle 470R 3W wirewound or metal oxide Fis SIRT Lee 1k carbon composition 0 25W ale 1k 3W wirewound or metal oxide Bee 12k5 metal film or wirewound 2W EW 20k 0 5W PRG eek A RE OE ee ne 300R metal film 0 125W Al e 100R 0 5W le 120R 0 125W ole 47R 0 5W EE 680k 0 5W alle EE 56k BW wirewound BAS 15k metal film or wirewound 1W NIE ee 200R 0 5W matched to 0 1 RORE et ee etree a cee eal E 4R7 0 25W AE E 270k 0 5W Controls Fil Zee le Neen ean E 47k 0 5W WE 100k stereo potentiometer audio taper le earner nea E 2 pole 4 or 5 position rotary switch Capacitors Power Supply Il KEE 1uF 630V polypropylene Capacitors Signal circuit two of each needed for stereo e IK KE 100uF 450V electrolytic C E re ees Not used jumpered see text Il E 47uF 400V EE 0 1uF 630V polypropylene Ir Not used Tr KEE 0 47uF 400V polypropylene 6106 C107 i eee 0 01uF 600V ceramic disc lU e 10uF 100V electrolytic Semiconductors and lamps Signal circuit two of each needed for stereo Semiconductors Power Supply UE EI De Red LED 1 7V D101 D102 D104 D105 wee UF4007 fast recovery rectifiers NEE DN2540N5 Depletion mode MOSFET HUE Zener diode 12V 1W NEE NE2 neon bulb D I EE TIP50A or equivalent EI nee ern te crt SP at Ee LM317 adjustable regulator Transformer Signal circuit two required DE Jensen JT11P 1 or equivalent 1 1 input Transformer and Miscellaneous Power supply transformer EE A
46. Book pre emphasis curve specifies time constants of 50uS and 15us The 3dB point for the high frequency boost is 3183Hz shelving at 10610Hz Maximum boost is 9 49dB at 20kHz www audioxpress com Only a handful of test CDs have been and make my own test CD To do so I needed a precise model of the Red boost in record is about 9 5dB To avoid clipping the 20Hz reference should be made with tracks for checking de em phasis in playback I have used Hi Fi News and Record Review Test CD II duced using CircuitMaker 2000 Figure 2 shows the simulation model T RC why that column is labeled LEVEL HFN15 which has pre emphasis tones at 1kHz 4kHz and 16kHz Chuck Hansen has used the CBS Test Disc _ by R1 R2 C1 and the 15us time constant is produced by R1 C1 Scal ing for VcVs1 is set to the value of R2 and IcVs1 is set to unity R3 is a load resistor for IcVs1 arbitrarily set to 100k tone The CD I made actually has two CBS 1 which has tones at 125Hz 1kH7z 4kHz 10kHz and 16kHz These discs are adequate for determining whether de emphasis has been imple mented but they don t have enough tones to give meaningful data about the accuracy of the de emphasis Some discs such as the Pierre Verany Digital cuitMaker 2000 like most schematic capture programs with simulation will pre emphasis applied according to the Test DV 788031 32 and
47. DC from the output I ll need 46 73V DC And the adjusting resistor will be 30 R15 46 73 00135 35 071Q The nearest E96 value is 34 8k Maximum current through the ad just string will be 1 42mA and max imum power dissipation in R15 will be 0 071W so a quarter watt resistor will be adequate The dropout voltage for a TL783 draw ing 43mA is roughly 5V making the usual allowance for regulator and resistor tolerances and my 0 25V fudge fac _ which the smallest specified EE will handle easily Pll begin with a 48V AC transformer _ BRIEF DIVERSIONS It might occur to you that the above calculations are the sort of thing that _ spreadsheets do well That s correct Pve designed a spreadsheet which does them but it s not quite ready for prime time _ yet Watch this space Speaking of spreadsheets I ve pro duced an Excel file with parts lists for the three examples above it s posted on the magazine s website www audioXpress com The lists include the power transformers and all the parts on the boards except mechanical bits insulating washers standoffs pins and tor the voltage at the regulator s input should be at least 57 0V DC C6 C15 470uF and C9 C18 33uF Doing the usual arithmetic the DC voltage on the first stage of the fil ter will be 58 2V DC under low line conditions Dropping this to 57 0V DC would require a 28 5Q res
48. STOM WORK audioXpress February 2009 29 assuming an ambient temperature of 40 C the chip temperature should be 77 C That should be fine the LM317T is voltage at the first cap is less than the The compressor will be attached to the regulator all the time it draws a mini rated to a chip temperature of 125 C mum current of 65mA This is enough to keep the regulator regulating happily so no additional load resistor is required PIL leave out R19 EXAMPLE 2 THE FILAMENTS What about using Supply 2 to produce 6 3V DC at 1 2A which will power four filter cap is 10 52V DC and the ripple is about 0 6V AC That s better doing the arithmetic for the dropping resistor R8 _ it should theoretically be 1 07Q 1 0Q is 200Q again then Radj 800Q nearest E24 is 806Q The regulator s dropout voltage from the datasheet is 2 25V so 12AX7s wired for 6 3V or two 6SN7s To avoid tedium I won t show most of my work this time Looking at the regulator if Ro a under worst case conditions the unregu lated voltage at the regulator s input needs to be mumble mumble 9 23V DC Using the same filter capacitors 1100uUF at the input 3400UF in stage 2 of the filter and a 12V AC wall wart transformer Vpk lo 11 57V DC ripple voltage will be 2 63V AC and the volt age at the first filter capacitor will be 6 99V DC TILT That s not going to work the
49. There is no need So if you estimate the total current to be dissipate 400 350 50 volts times by the pass transistor so it should have some heatsinking The output voltage is set by 1 25 times the ratio of R108 to R109 R108 dissipates a bit over 2W so size it gener ously It may be useful to make it from a series or parallel combination of two 2W _ days so why not use them R101 and R102 keep the ripple currents under resistors The output capacitor C104 has a re sistor in series with it basically to flatten out the inductive output impedance of the LM317 regulator This series resistor means that money spent on an ultra low ESR cap for C104 is better directed toward booze and gambling The signal _ the resistors are reflective of RMS ripple currents Wirewound resistors are pre circuitry is locally bypassed by C5 How do you get the 400V for the I found a nice power transformer from son why a surplus unit won t work just as 400V at reasonably low ripple to feed the a conventional E I core other construc couple that noise into the power supply High speed rectifiers are cheap these control and give the diodes some pad ding during power up I chose the value of C101 to give 400V at the raw supply output its relatively small value again keeps ripple currents low This is fol lowed by two RC filters the ratings of ferred but the metal oxide type will also On the primar
50. a sweat not escaped my notice that this preamp looks very much like the input stage and REFERENCES phase splitter for a push pull tube power amp And there s no reason not to build a tube unity gain power amp with a bal more than a fair share of construction work I thank Nelson Pass for the use of Fig 1 and for his support and encour agement Morgan Jones for his usual perceptive critique and the CV1988 and Cynthia Wenslow of Rising Moon 1 Morgan Jones Valve Amplifiers 3rd Edition Studio and Steve Eddy of Q Au dio for assistance with drawings and organization aX The book on loudspeaker testing from MTM inventor Joseph D Appolito To order call 1 888 924 9465 or online at www audioXpress com Old Colony Sound Laboratory POBox 876 Peterborough NH 03458 0876 USA Learn to test your loudspeakers with this great tutorial More than simply a how4o the book uses specific examples to demonstrate the principles involved The final two chapters provide a highly readable explanation of the theories needed to understand PC based electrical and acoustical data acquisition and analysis systems Examples of measurements made a ig and CLO are included A musthave book r experienced hobbyists speaker designers and technicians 1998 176pp 82 x 11 softbound ISBN 1 882580 17 6 Sh wt 2 Ibs eeee eee eee eeeeeeeeeeeeeee eee eeee Shipping and handling additional Tes
51. across R4 this biases the sources more positive This is equivalent to bias ing the gates more negative regarding the sources more negative Vgs And this decreases the current which has the JFET in Fig 1 Low Noise Ch Measurement Range Test Circuit Preamp i with Averaging Function Wa Sr RMS Probes Grounds Not Used AC Line Ground invert AC Line FIGURE 2 Noise test setup Ground A 2933 2 FREE SAMPLE IF YOU BUY SELL OR COLLECT OLD RADIOS YOU NEED ANTIQUE RADIO CLASSIFIED Antique Radio s Largest Monthly Magazine 68 Page Issues Classifieds Ads for Parts amp Services Articles Auction Prices Meet amp Flea Market Info Also Early TV Art Deco Audio Ham Equip Books Telegraph 40 s amp 50 s Radios A More Free 20 word ad each month for subscribers Subscriptions 24 for 6 month trial 45 for 1 year 60 for 1st Class Mail Call or write for foreign rates The Collector s Guide to Antique Radios 7th Edition by John Slusser amp Radio Daze 10 000 prices 1 200 color photos 24 95 Payment required with order Free shipping In Mass add 5 tax A R C P O Box 802 A16 Carlisle MA 01741 Toll free 866 371 0512 Fax 978 371 7129 Web www antiqueradio com E mail ARC antiqueradio com audioXpress February 2009 25 effect of opposing the initial assumed preamp In the low noise measurement preamp the FET sources after the 1
52. an relying on the simple basic SPICE default perfect in ductor LOUDSPEAKERS The usual published equivalent circuit impedance and phase angle parameters to be displayed on screen or used to gether with other components in SPICE pacitor or a complete speaker system for a loudspeaker driver has been de vised to provide simulations only at au dible frequencies Such models then as sume the speaker driver continues to act as a perfect inductor so they do not include the inevitable self capacitance As a result these models assume that matics using SPICE At audible fre quencies with modest component val ues that can work quite well however at higher frequencies every component speaker cable aX speaker impedance continually increases with frequency Unfortunately that ever increasing impedance is not found when measuring actual speaker drivers even at used whether inductor resistor capaci tor and especially speaker drivers must use complex multicomponent models to accurately match resonant frequen _ cies Every inductor or speaker voice coil includes significant self capacitance and resonant frequency peaks and troughs Simplistic SPICE simulation of an in ductor shows impedance continually nated with 570Q at 865kHz then a low of 75Q at 3 5MHz a small 108Q peak at 5 85MHz to another low of 72 5Q at 9 5MHz I also tested
53. ate each tone in Table 1 for Sony Volume Pre Emphasis Test CD V2 wav a length of 30 seconds at a level of exactly 10 0dB for eee a total of 28 tracks I put 4 seconds of silence at the end of each tone plus 1 minute of silence at the end of the last track Then I marked the entire file and performed a simple copy and paste du plicating all 28 tracks again These duplicated tracks 29 through 56 are the ones that will have pre emphasis added Sound Forge has two different Vol CPU volume by any level you choose The which is a bit too coarse for the lowest frequencies There s another Volume plug in under the FX Favorites menu that allows adjustments in increments of 0 001dB which is the one I used Fig boost that tone by the amount indicated ELECTRA PRINT AUDIO CO AUDIO TRANSFORMERS Single Ended e Push Pull e Parafeed e Cathode Follower e Interstage e Line Level Outputs e Audio Chokes e Moving Coil e Step up down e Low level input e Phase splitting e Silver windings e Nickel core designs POWER TRANSFORMERS e High Voltage e Filament e Filter Chokes Custom transformers built to your specifications Custom Amps and Pre amps of our design Visa MC amp Amex ELECTRA PRINT AUDIO COMPANY 4117 Roxanne Dr Las Vegas NV 89108 702 396 4909 Fax 702 396 4910 electaudio cox net www electra print com 34 audioXpress 2 09 Untitled Volume
54. ate peak value and the LDR divides down the volume so that the deter mined peak value is not exceeded R9 was not used R10 can be used to adjust the time constant of the circuit Some information about compression and limiting can be read at sound westhost com compression htm The Audiohm couplers use the CdS cell like mine does set the high end rolloff _ for my particular application suit yourself for yours The basic 3dB bandwidth of the amps is around 1 9MHz which for any gain value used here would provide sufficient bandwidth for the audio range Read the ADJUSTMENTS section for how to adjust the low and high frequency rolloff Note that use of the LM386 is en tirely optional you can connect any amp of your desire in place of it externally can find no need for caps across R14 and or R2 _ HELP WANTED I need help finding a source for bi tuminized felt If you remember Leak and others used it in their speakers to reduce wall vibrations I have used layers of shingles which work OK but would prefer the right materials Vincent Mogavero VMogavero aol com aX Continued from p 39 but briefly presented chapters the snare drum and the piano The discussion of control and imaging here should be of interest or perhaps controversy to most readers The specifics studio The final chapter is on mixing down the multi track recording in basic terms
55. ator wanted to see Vunreg was 18 25V DC So 17 Vdrop2 V DC lo Vunreg 19 18 1629 0 95 18 R7 Vdrop2 Itotal 0 93 0 257 3 602Q The first choice would seem to be a 3 6Q resistor but what about its tolerance If it s 5 high that s 3 78Q the voltage drop will be 0 971V That would make Vunreg 18 21V Hmm the difference For the raw supply design PI start with the transformer choosing one with 18V AC That s the voltage of each winding tor will be ings for the two halves of the supply is only 0 05V under worst case condi tions I think it s safe to specify a 3 6Q 5 resistor the power through that resis 19 P Itotal R7 0 257 3 6 0 238W 2W Being cautious I ll go for 2W its cutoff frequency Fc is 13 0Hz The ripple will be mainly 120Hz on a 60Hz electrical grid anyway and 26 Vregdrop Vunreg hi Vreg lo will be reduced by approximately Fc Fripple 13 0 120 0 108x m Since the ripple at the first capacitors was 0 561V AC the ripple on the 27 P Vregdrop Itotal 9 04 0 257 22 Vripple2 0 561 0 108 0 061V AC With a Wakefield 637 10ABP heatsink the thermal rise should be about 37 C second capacitors will be e That s plenty low enough to ignore AUDIO TRANSFORMERS AMERICA S PREMIER COIL WINDER Engineering e Rewinding e Prototypes DISSIPATION IS RUINATION The las
56. ce which affects the resonance width together with a large shunt loss resistance value which limits the peak impedance value These series and shunt resistances are best devised by running a few simula OH a 2 ee eee Au t 0 00K o NO inducti Op s air core induct tions adjusting values as needed to ap proximate measured values I find that a makes a good starting value The mea INDUCTOR CIR zm mm mm ms bm mm rm zm ze ee I mm mm mm mm wm wm wm wm e e db wn wm wm wm wm mm mm mm zm mm V_inductI R9 used measured 843Q at 500kHz and the _ T27 tweeter measured 825Q at 2 4MHz Within this frequency band other driv ere all measured as intermediate imped ance peak values With inductors typically peaking around 40kQ clearly the com _ bined impedance of crossover and speaker driver at high frequency must approxi mate 5 600Q when resonant reducing with further increase in frequency I had two quite different crossover 5 and 6 which measured a peak imped ance of 525Q at 900kHz The other the evaluations paper measured a peak im pedance of 575Q at 1 5MHz At frequencies higher than this induc tive resonant peak the inductive leading phase no longer applies The measured phase angles now lag and the inductor like a capacitor This often results in the circuit drawing unexpectedly high and capacitive currents MODEL DEVELOPMENT The simplistic idea
57. ce far better more realistic simulations closely matching actual measured values for both capacitors and inductors ESP replica crossover measured Direct impedance and phase No amplifier zobel or cable used Impedance magnitude and phase measured Frequency checked using freq counters define Listal 0 0 0 1000 4 89 11 3 2000 10 61 19 2300 13 1 12 2470 13 49 1 2700 13 2 7 3000 10 77 12 6 F Out1 v Out 3 F 4000 5 48 10 6 5000 3 92 0 8 6000 4 01 9 3 7000 4 51 15 6 8000 5 20 20 4 Listval 9000 5 86 24 4 a Out 3 10000 6 58 28 8 11000 7 18 31 12000 7 83 33 5 13000 8 41 35 2 14000 8 98 36 6 15000 9 56 38 16000 10 11 39 6 17000 10 66 40 8 Frequency Domain Analysis ONLY 18000 11 20 42 19000 11 75 43 20000 12 31 A4 FIGURE 9 SPICE One Port or Z_ block model subcircuit which you can use in simulations ESP replica crossover no cable no zobel measured direct Out a In 1 Z Block Impedance_Mag Phase 22 audioXpress 2 09 www audioXpress com rough guesstimate model would be far more representative of a real inductor _ pedance typically around 100Q None _ of these drivers measured as a particu larly high impedance following this first resonant peak The Z_block model Fig 9 allows a CSV listing of measured frequency power amplifier was series resonant at at high frequency th
58. ch drop about 120V and you want 100V or so across the tube to en sure proper operation at the high voltage swings required R11 and R12 should be very tightly matched so that the balance and distor tion are optimized If you buy a dozen or two 1 resistors it should be easy to find a couple of pairs that match to within 0 1 the absolute value isn t critical the relative values are so the calibration of your ohmmeter isn t Voltmeter going to be the determining factor in matching With the cathode resistor at 120V the grid will need to be slightly nega tive of that value about 118V or so This presents a small problem the 6SN7 FIGURE 7 Circuit to test operating current will likely have 170 180V or more on PRAXIS audio measurement system e Provides full featured 3 Primary Plot praxis BEIS e Measurement Guide BIS EI f Format File Print Clone Hold ClearMulti Table Fam 7 5 customizable measurement Diagem Bint Pianis Hep Ent Polar px2 Frequency Response acq 7 5 2002 1 35 51 PM including loudspeakers room acoustics transducers AR Us GC j Frequency 6 75kHz EEN i dB div 2 IN electronics vibration testing PA pin alignment and quality control i ae d Cantar 10 4940 EECH d I roiorz Flexible portable system rae Format File Print Clone ClearMulti Table advances with the latest N de QA_BassRestored px2 acq 5 18 2003 10 26 45 PM I
59. critical than the lower value load resistors and power amp input make a supply that is orders of magni tude better than what you strictly need for not much more complication and cost than a supply that just does the minimum So I will don my 16 piece of flair and use an active regulated sup ply with reasonably low noise and source impedance The Maida regulator based on Na tional Semiconductor s LM317 has as Joe Curcio s Stereo 70 conversion own Red Light District Because of Che signal levels involved and the in herent power supply rejection of the signal circuitry I ve used a simplified version closer to Curcio s than Jones s or the original Maida circuit If you want to go crazy and add all the extra bypass caps and the protection diodes needed to deal with the consequences vou can boost the regulator s measured performance but the sonic effect will be minimal to nonexistent The Maida regulator is easy low cost quite stable the voltage amplifier and 100 local degeneration of the output stage go a Jong way to ease the demands for ultra _ low impedance Nonetheless it is easy to and reliable Because the design is discussed thor oughly in Maida s paper I wont rehash it in detail I used a TIP50A as the Parts List Resistors Signal circuit two of each needed for stereo Resistors Power Supply FRE eset haat CR NEA Pe th ee 18k metal film see text AER
60. d with the 50S 15uUs Red Book pre emphasis curve www audioxpress com gions List and change the markers to re once CD authoring program that pro editing CD Architect makes tracks from Fig 7 You should see all 56 tracks you ve produced in Sound Forge On the right you ll see a column called Emph there s a box for each track all un checked Check each box from track 29 through track 56 This will tell CD Architect to write the pre emphasis flag for each of those tracks This flag tells your CD player or outboard DAC to turn on the de emphasis circuit It s important to understand that there are two processes involved in pro ducing a CD or individual CD tracks with pre emphasis The first step is to apply the correct high frequency boost to your WAV file I did this with Sound Forge one track at a time adjusting the level of each tone according to my circuit simulation But altering the frequency response File Eat wm Specie Protas Eiter Took Fi F asier Option Windew He D De y TK EK e SO F Go FF UU SM We H Set mo Lengel gatai H H OH DORE EEN CO ee i Oe 07 tet ae Ierwe OOS DBL Dei Ho DOG Lem Deo Dei eu DOG SELF COOC IL elie 4 EN DO 24 081 DORA CG DI I ent CO SEL DC CC CL 00 08 4 0 Dro Gre 11 Cee a Dr 081 Gedo TT ug DN GSO Cen CEO Der CH San Do 4 007 DAE Oft OE AEF OO eds hih H00 OTT Dib DO ds 00h oo CT DO Od 0
61. deal capacitor exhibits a 90 phase angle between its applied voltage and through current but every practical capacitor must use metallic _ frequency the metallic electrode and con nection resistances dominate the dielectric contribution ESR then reaches a mini mum value after which it slowly increases due to skin effects It is certainly never a constant as many wish to believe 90 and increasingly so with an increase ponent capacitor model shown in Fig 7 _ can suffice over a narrow frequency range _ both for AC and transient simulations placed across a perfect capacitor results in electrodes and metallic external connec tions both of which introduce resistance and inductance Any solid dielectric in sulator exhibits its own dielectric loss es Combined together these result in a phase angle at low frequency less than in frequency This degraded phase angle can be represented by using the appropri ate very high value shunt resistor which the same phase angle This high resistance value is usually measured as conductance in siemens the reciprocal of ohms This degraded phase angle is more usu ally represented by the equivalent very low value resistor in series with the perfect capacitor This ESR value or equivalent series resistance can also be derived from measured values for capacitance and tand by multiplying tand by Xc the capacitive reactance at that specific freq
62. dels un less their components are specifically in tended for use at high frequencies for example Coilcraft which provides ac curate simulation models for use in the Microwave Office software As a result simulations of larger value inductance modest frequency result in large errors and misinterpretation of circuit behavior Let me examine actual measurements of a few real inductors Fig 1 All practical inductors include self capacitance between adjacent wire turns and from the start and finish end to end turns When using toroid cores these end to end turn capacitances dominate wound space typically about 90 between start and finish winds Next I examine typical SPICE plots using the default inductor models provided as standard 2 The SPICE default inductor model Fig 2 agrees with my measured val ues at audible frequencies and contin ues to agree up to 100kHz However by 200kHz I find the 1 156mH induc tor simulation suggests an impedance of 14540 whereas the measured value was 1779Q Clearly these errors then increase _ rapidly with frequency Many writers have used SPICE default models to simulate amplifier Nyquist response up to 1MHz _ driving into a simulated loudspeaker cable and simulated crossover speaker loads _ using only basic SPICE models for cable and speaker with invalid unreliable con clusions Eve
63. e first presents the cold hard equations even if it is the type of math that I would not ex pect to see until after at least one or two calculus courses Then he presents the explanation in language and images that allow the beginner to grasp the concept The book also assumes that the read er has knowledge of how things really work and so skips what many basic books would address For example he doesn t mention microphones until chapter 13 He assumes that once you know how air is modulated in chapter one you can fig of chapter two This undoubtedly is because Case is re ally teaching this subject He understands sional audio equipment That three chapters about the basics The _ first how air is transmitted by sound waves is represented graphically and use the processes are clearly based upon the author s actual experience This gives thor then details how this is handled in de mixer and multi track recording pro _ cess Concluding the introduction section ENTERING THE STUDIO is a chapter packed with information on how we perceive these sounds and can use this to advantage This is the best the explanation of what a volume control _ is and how to use it to be much earlier in the presentation It begins with an ex planation of the types of controls and how they work This material contains a _ brief explanation of how to mix music in a contemporary pop
64. e will be US 24 per unit there ll be a small discount for quanti ties gt 2 plus shipping handling You can e mail me at pstamler pobox com and no I don t expect to get rich from this project either Many thanks to Cory King of the IwThe21st Century Analog playback should be More Technical More Objective More Affordable The Technics SL 1200 SE Five Base Models Seven Genuine Upgrades Starting At 475 VI Fhe Cudiaphile Custom KAB Preserving The Sounds Of A Lifetime www kabusa com audioXpress February 2009 31 sound solutions By Gary Galo Regular Contributor A De Emphasis Test CD You ll find this test CD more useful than the existing published versions hen the compact disc was and Philips built an op tional treble pre emphasis phasis standard has often been referred to as a noise reduction system but this outboard D A converters have stopped tion True the high frequency boost in the Monarchy M24 I reviewed in the curve into the Red Book specifications for the format Initially the CD was intended to be a 14 bit medium which pushed the limits of storage and signal processing at that time By the time the CD was actually introduced to the public in 1982 the resolution had been increased to 16 bits Recordings made poor linearity at low signal levels par even the 16 bit converters in the 1980s had shortcomings in this regard The Red Book p
65. em Figs the equation l O n FAL derived from the more common equation E 2r LC Using the 400kHz resonance for the ments take little time but how well do _ they work Figs 3 and 4 With a few extra components you can add an additional node to also model _ ESP replica assembly as used for my cable 1 156mH inductor Fig 1 you find its equivalent capacitance the combined ef fect of its turn to turn and end to end capacitances to resonate at 427kHz must secondary resonances Loudspeaker drivers can be treated as an inductive component but for accuracy to high frequency they require several many additional nodes With the im pedance of a driver measured to about 10MHz in addition to the usual audible impedance reduces with frequency just be 120pF Using just these values results in too sharp a resonance because the real inductor exhibits a measurable DC re sistance of 2 184Q almost all is effec tively in series with the inductance Also because the resonating capacitances are complex they will exhibit resistive losses _ frequency resonances which are usual Je well modeled the drivers I measured ranged from an 18 250W Goodmans _ Power series with a peak impedance of 495Q at 400kHz to a Kef B110 with peak impedance of 570Q at 1 1MHz In my ESP_replica network the bass driver _ some loss will be in series with this capac itan
66. er bonus of CCS loading of the rejection Any noise on the B rail is CCS source impedance The low imped ance of the LED and the high imped CCS output impedance any power sup ply noise or instability is reduced by over 100dB That s a very nice way of remov ing power supply criticality The outline of the voltage gain stage is shown in Fig 4 One notable feature is the use of an input transformer In an earlier article I justified its use in some detail basically role As you modulate signal across the CCS the drain source voltage varies _ Though the curves are quite flat i e the balanced or single ended drive A 1 1 unit it provides outstanding common mode hum immunity and galvanic isolation while at the same time facilitating either is appropriate here The gold standard is the Jensen JT11P 1 which I used in the prototype but other units from Cinemag Sowter and Edcor should work equally well All of these have very low distortion and excellent bandwidth CONSTANT CRAVING A very good choice for a con stant current source CCS is a cascode and among cascodes one of the best performers is a FET cascode Finding JFETs that will withstand the voltages and cur rents involved is not trivial High voltage MOSFETs are plentiful and in cascode can make an ex cellent CCS Unfortunately the biasing arrangements necessary for enhancement mode devices can complicate things Th
67. ere is relief at hand de pletion mode MOSFETs are becoming easier to get and the DN2540 is cheap available and works perfectly in this service A simple yet high performance CCS using these wonderful de vices is shown in Fig 5 This two terminal self biasing cascode CCS with a source impedance in the 1000M range can withstand 8 audioXpress 2 09 2 400V without breaking a sweat To understand the operation of this cascode CCS first consider Q1 Its cur plate and LED biasing of the cathode is the greatly increased power supply _ fixed Because to first order the gate cur divided down at the output of the 6SN7 by about the ratio of plate resistance to rent is set by the value of R5 For a fixed current the voltage drop across R5 is rent is zero the drop across R6 the gate stopper put there to prevent oscillation is zero so Q1 s gate to source voltage is fixed at I times R5 ance of the CCS mean that this ratio will be very small For example with a 7500Q plate resistance and a 1000M 3 FET characteristic curves that means that the current through the FET is The second FET has a more subtle output resistance is very high they are not perfectly so Additionally the drain source capacitance is significant espe ro find two devices with the same part number having I or g values that differ by a factor of three I bought sev eral tubes of the DN2540 and kept
68. es that you can as 6SN7 Medium mu triodes also have the ad 300pF This might upset some driving sources but these days that would be a minority Still it is something to con sider especially when choosing the value source impedance driving the volume control the worst case source imped ance is at the halfway 6dB setting and je about half the volume control total resistance This impedance combined with the tube s input capacitance forms a pole which will often impinge on the audio band A 10k potentiometer combined with a D3a input will give a 3dB point of about 100k which is a satisfacto Ground FIGURE 3 Simple common cathode voltage amp feeding a split load inverter www audioxpress com ry bandwidth but higher values could begin to be problematic Gain is also significantly higher than target so some careful reduction measures would need to be designed in Fortunately a very common medium mu triode the 65N7 has demonstra bly excellent linearity at these voltage swings the requisite gain sufficiently low input capacitance 80pF and is more than the desired 20V peak out put because the phase splitter will have slightly under unity gain It will be light efit of visually indicating that the tube is _ indeed drawing current widely and easily available in an impres sive variety of flavors Distortion perfor mance is comparable to the
69. f the folks who work a mixing console for The next section of the book is on what _ he terms amplitude effects The presenta tion in the fourth chapter on distortion is the start of this title material The chapter is too short from an audiophile s point of view but does offer insight into the re cording artist s deliberate use of distortion for musical effect and some of the prob lems to try to avoid A nice feature here and throughout the book is the listing of specific musical selections to illustrate the written examples ure out how this gets into the signal flow fun or profit Chapter 11 on reverb is aimed square ly at recording technology devices and techniques But it will give most readers an insight into what the still emerging digital technology will do to live perfor mances and the performing space Section four which covers the basics of mixing is where this work really becomes a more complete studio book It explains very well why there are both pre fader and post fader sends on mixing boards The basics of where to assign which sig _ nal and how to get started are reasonably Continued on p 45 audioXpress February 2009 39 XPRESSMail CROSSOVER APPROACHES I want to thank G R Koonce for his recent article Passive Crossover Lin This topic is of interest to me and it is noteworthy that your conclusions and sults in a lot of de
70. gnet The tweeter is rolled off on the low end with a Fostex Tin amp Copper foil capacitor The system frequency response is 55Hz to 35kHz at 95dB The BK 16 cabinet is made from Baltic Birch plywood and is sold flat unassembled unsanded and unfinished Cabinet dimensions are 9 75 W x 14 75 D x 29 T The kit includes e Pair of Cabinets Flat Pair of FF165K full range Pair of T90A horn tweeter Pair of DB Cup Input cup Pair of Crossovers Nordost 2 Flat wire for tweeter Instructions Kit Price 750 00 pair Parts without cabinets 454 00 pair Cabinets only 148 00 each AN MADISOUND SPEAKER COMPONENTS INC Jr 8608 UNIVERSITY GREEN P O BOX 44283 MADISON WI 53744 4283 U S A TEL 608 831 3433 FAX 608 831 3771 e mail info madisound com Web Page http www madisound com audioXpress February 2009 17 a knowledge of an inductor s resonant frequency you can easily devise a more peak impedance shunt resistance of 40kQ appropriate computer model You can cal culate the effective self capacitance using sured DCR in series with the inductance less about 0 10 for those lead wires which De outside the main winding provides an excellent fit to measured slope values _ which can be finally tuned by adjusting a small resistance in series with the shunt capacitance These cut and try adjust speaker cabinets available The first was my horn loaded two way syst
71. have for fans of early sound recording and music history Sh wt 1 Ib CDME1 21 95 www audioxpress com A Matter of Coincidence recorded by Brian H Preston Great collection of music recorded using a crossed microphone tech nique Includes vocal chamber organ and jazz selections A great demo disc for your audio system 2003 Sh wt 1 Ib CDBPI1 ponent is then better described using a would exhibit near constant degraded phase angle or tan with frequency Be cause for each doubling in frequency ca pacitive reactance halves the ESR for the ment frequency SPICE allows model values to be modified using the model statement for transient or frequency F ing for each octave increase in frequency When using the shunt resistor or G its equivalent conductance both values are strongly frequency dependent In this case to maintain this degraded measured phase angle the shunt resis INDUCTORS tance must more than halve and conduc tance G must more than double for each can quickly calculate the values needed for an inductor model Unfortunately for distributed transmission line like model To complicate matters further in real components the values of these parasitic elements invariably change with measure factors for AC simulations but these two options are mutually exclusive CAPACITORS Begin by examining an improved capaci tor model Fig 7 The i
72. he input with HL I LAL d GD 100 200 600 1k 2k ek 10k 20k DJB Ren geg SEL EE eS ady kp d JEE KEE 1 I 1 1 4 bodd do bow 104 Bul 204 LH 20 60 100 200 E00 1k 2k 6k 10k 20k j T 1 Lt COOL 60100 200 600 ik 2k 5k 10k 20k FIGURE 1 Performance of two way minimum phase crossover with ideal WTW array 40 audioXpress 2 09 the supplement being the other CO the high pass However I ve worked at _ making a good supplementary CO for ear Phase Speakers June and July 08 years now with little success the major problem being that the supplement is a thin felt rings on the tweeter faceplate virtu low order CO often with ears This re of their measured limitations I appreciate your good work and look nel It does not sound as though Mr Craw ford has had much luck with this approach Dick Crawford 2 dickcrawford1934 aol com crossover approach he had developed With 3 his permission will document it here for would like to thank Mr Crawford for his kind comments on my article and look for _ ward to reading his article Tweeter setback _ order low pass LP with a Q of 0 5 at twice the intended crossover frequency and pair it _ tweeter location increased the distance be tween the tweeter and the grille cloth After G R Koonce responds generally reduces time dispersion bu
73. in the olds related to levels and frequency but it bears no reference as to where this infor mation came from This book hits its target reader dead on but goes way beyond that At 39 95 this book is a very low cost way to get a seat in the first class compartment on the audio clue train aX The Newest Products and Technologies are Only a Click Hh i a 7s pn TS te nena Over A Million Products Online More Than 366 Manufacturers Easy Online Ordering No Minimum Order Fast Delivery Same day Shipping 800 346 6873 The Newest Products for Your Newest Designs Mouser and Mouser Electronics are registered trademarks of Mouser Electronics Inc Other products logos and company names mentioned herein may be trademarks of their respective owners vy tubeampdoqeer ca To eee TUBE ANP DOCTOR GmbH 675571 Worms Germany E Weinbrennerstrabe 1 7 fa phone 49 0 6247 90405 0 fax 49 0 6247 90409 28 audioXpress February 2009 45 E a Super FI www audaud com Claviers Mozartiens Pierre Goy vis vis clavichord piano carr Lyrinx SACD LYR 2251 3 t SI ry maza rtiens between the four keyboard instruments French Romantic Organ Music arranged for Brass Quintet amp Organ Guilmant Vierne Lefebure Wely Boellmann Elmar Lehnen Seifert Organ Kevelaer International Brass Quintet Audite MC SACD 92 556 The organ of the papal Basilica of St
74. int as the RC product of R3 problems am not against negative feedback but if the electronic work is very carefully done with no noise and good acoustic performance why do it Anyway the amp has been working well from about one year in one of my friends home VOLUME CONTROLLER I m not quite sure what to make of An Automated Level Control in your Oc tober issue p 18 The author states that It does not compress the peaks but simply turns down the volume as you would do with your volume control I fail to see the distinction between this circuit and any active compression cir CONTRIBUTORS simple and have much improved audio charac teristics over CdS You can see a similar compressor circuit in their Technical fine for ham operators I guess However capacitors across their feedback resis Stuart Yaniger The Impasse Preamplifier p 5 resides in Benicia Calif Cyril Bateman Simulating Inductors and Networks p 16 has done extensive work on capacitors and resides in Norfolk England Dennis Colin Noise Measurements of the LSK389B Dual JFET p 24 graduated with a BSEE from the University of Lowell MA and is currently an Analog Circuit Design Consultant for microwave radios Previously a band keyboardist and a recording engineer he has been published in the Journa of the Audio Engineering Society Colin has demonst
75. istor the next lowest value is 28 0Q That ll dissipate 0 055W a quarter watt resistor will do The nearest E24 value is 27Q The rest of the design is pretty straightforward maximum voltage under high line conditions would be 72 1Vpk and it s prudent to add another 10 fudge factor for a lightly loaded trans former That means the maximum volt age on the first capacitor could be about 79V DC so the caps should be rated at 7 100V working voltage With the capaci What about idle current Unlike the regulator should keep regulating under microphone connected Texas Instruments specifies a mini mum current draw of 15mA for the TL783 to perform properly Under worst case conditions Vref 1 2V DC R18 1010Q the adjustment string will draw 1 27mA to draw 15mA total the load resistor R21 will need to pull an additional 13 7mA The worst case low regulated voltage will be 45 0V DC so a 3278Q resistor will do the job _ The next lowest E24 value is 3k which Webster University Audio Construction Club I hope you have as much fun using this board as I had designing it aX If the regulator is at its highest possible voltage 50 1V DC and the resistor is at the low limits of its tolerance 2850Q then 21 Pdiss 50 1 2850 0 88W 2 A 2W resistor is the minimum for reliable performance and since there s room on the board for a 3W Pana
76. kHz 3 54 5kHz 4 53 6kHz 5 38 7kHz 6 1 8kHz 6 69 9kHz 119 10kHz 7 61 11kHz 7 95 12kHz 8 25 13kHz 8 49 14kHz 8 71 15kHz 8 89 16kHz 9 05 17kHz 9 18 18kHz Bra 19kHz 9 4 20kHz 9 49 Ja x Ca TTT TTT a Bypass Real time More Occasionally I found that the tones were not at the level they should be In this case I highlighted the entire tone and normalized it to a level of 10 0dB then re peated the above procedures It always worked on the sec ond attempt Sound Forge 9 0 comes with a Mastering Equalizer plug in made by iZotope which includes pre empha sis and de emphasis curves However the algorithms dont seem to have the precision I get from manu ally adjusting each tone according to the simulation As an example Table 1 column 3 says that 11kHz should be at 2 05dB if I pre emphasize all 28 tracks using iZotope s plug in 11kHz is LEVEL REF TO Tracks OdB 10 000 1 29 9 999 2 30 9 996 3 31 9 904 4 32 9 862 5 33 9 814 6 34 9 759 9 9 697 8 36 9 630 9 37 8 710 10 38 7 570 11 39 6 460 12 40 5 470 13 A 4 620 14 42 3 900 15 43 3 310 16 44 2 810 17 45 2 390 18 46 2 050 19 47 1 750 20 48 1 510 21 49 1 290 22 50 1 110 23 51 0 950 24 52 0 820 Us DS 0 700 26 54 0 600 21 55 0 510 28 56 Pre emphasis levels based on the circuit simulation of Figs 1 and 2 The test CD has 28 tracks recorded flat at 10dB and 28 more recorde
77. l with a 15k resistor shunting it no ca There is some minor variation depend ing on the position of the volume con trol this is because of the 80pF Miller input capacitance of the 6SN7 but it is still pretty close to optimal at all posi tions For other transformers or other values of volume control either follow the manufacturers recommendation or adjust the values of R1 and C1 using a square wave source OUT HERE IN THE FIELDS With the voltage amplifier stage T2 D101 R101 eh R102 D102 C106 p C107 Y HTR CT FIGURE 6 Complete schematic 10 audioXpress 2 09 wrapped up you can move on to the output stage which is tasked with taking a single ended signal and converting it ro balanced the so called phase splitter pacitor needed works nearly perfectly ated interconnects The ECC88 6DJ8 in the output _ stage offers a very low source impedance 100R typical and it is an easy tube to find in all sorts of flavors Because of 3 the 100 degeneration in this circuit there s little advantage to using any of the collectors items in this spot And other tubes will also work well here an ECC99 for example would be a fine choice I have substituted 7308 6KN8 and 6922 with no measurable or audible difference The key virtues are reason then driving the power amp and associ R2 c109 To HTR GND CT T R
78. ld capacitors resistors and especially induc tors measure quite different from their SPICE predictions ing they do not however represent any All components include parasitic ele ments so are better described at least to moderate frequencies by combining all three elements capacitance resistance and inductance for each device For realis tic high frequency simulation every com V Cap Mapiin V Cap wii EGh V Cap eimple V_induct_complex Inductors Y induct R58 simple Complex R65 0 001 ar a Le With ESR and DC leakage Capacitors Simple complex R66 L41 5 R74 0 02 C35 60ue R69 0 003 1 04u Maplin with 4 5 leadwires and resonance FIGURE 7 Simple and complex capacitor and inductor models olon SOUND LA JUYVINY Music Olan amp D Great Opportunities to Audition your Audio System Order these music CDs by calling 1 888 924 9465 or visit www audioXpress com to order on line Old Colony Sound Lab PO Box 876 Peterborough NH 03458 0876 USA Toll free 888 924 9465 Phone 603 924 9464 Fax 603 924 9467 E mail custserv audioXpress com www audioXpress com ZU audioXpress 2 09 Sounds Cylindrical produced by Joe Pengelly A unique collection of music recorded on cylinder between 1900 and 1929 Twenty one tracks containing a variety of musical selections including an early soundtrack recorded in 1913 A must
79. lized models for ca zm mm mm mm mm mm mm mm mm mm mm mm mm mm zf zm mm mm mm mm mm mm mm mm mm mm mm mm mm zm mm bb zs mm mm mm mm mm mm mm mm mm mm mm mm mm mm mm FIGURE 3 Simulations which closely approximate my measured impedance values for the first or main resonances of the 1 156mH also 250uH Falcon Electronics air core inductors shown in Fig 1 18 audioXpress 2 09 www audioxpress com R61 1 v induct1 V induct DC OAC 2 FIGURE 4 The simple SPICE models as used for the Fig 3 realistic simulations of the first or main resonances for the 250uH and 1 156mH Falcon Electronics inductors a ka d ss zm sms zm zs bs ee eee shee bebe zs ss zs sl zs ss ss bs D D i i E 0 D D D i D bed 10K 100K T k F FIGURE 5 Simulated impedance of 3 5mH ferrite cored inductor as used in my horn loaded speaker crossover network Actual measured peak impedance was 40kQ at 210kHz YV induct e CH SHOP ONLINE OR ASK D FOR A FREE CDROM Lef CATALOG SOLEN cs 4470 Avenue Thibault St Hubert QC J3Y 7TS DC o AC 2 D Canada i Tel 450 656 2759 3 5mH Ferrite core Fax 450 443 4949 Email solen solen ca FIGURE 6 Model used for Fig 5 simulation Web www sesolen ca audioXpress February 2009 19 pacitors resistors inductors and trans mission lines provided in most SPICE based simulators can be useful for DC transient and low frequency AC model practical components In the real wor
80. llied 6K3VG 650VCT at 40mA 6 3VCT at 2A NEU H Power Entry Module SAE PM1B 6 included with DIY 1712 case 12 audioXpress 2 09 www audioxpress com pass device but there s no reason why any other moderate power transistor wouldn t work there A power MOS FET could also usefully serve in that position Under operation assuming a 400V raw supply a good compromise between headroom and pass transistor dissipa _ raw supply Looking in the iron pantry do well Allied the 6K3VG rated at 650VCT 40mA It s overkill for this job but it s reasonably cheap 25 has a shield layer and is easy to find There s no rea approach you take make sure that the tion the regulator needs only to drop 50V but remembering that the regula tor will be turned on and off with the preamp do not try using a low voltage device for the pass transistor The regu lator draws about 6mA the signal cir cuits for a stereo version another 32mA _ well All you ask is that you end up with 40mA the pass transistor and LM317 _ regulator This particular transformer has 0 040A or 2W Most of that is dropped tions will also work some even better _ but avoid toroids With the universal _ adoption of compact fluorescent lights _ and switching supplies power lines are _ noisier than ever and toroids diligently shortcut and used the 6 3V winding on the power transformer
81. llow er the plate should be at AC ground Conveniently there s already a capacitor there C3 So by connect ing the output side of C3 to ground you have by passed the plate to ground and made a single ended output available at C4 This is most easily done by making up a set of in terconnects with XLRs at one end and RCA plug at the other Inside the XLR wire a jumper between the Ou Ces e plate output pin and ground That way you can go back and forth between bal anced and single ended just by plugging inverts polarity I could hope for Mid band THD was into a 10k load single ended This broke down as 68dB second harmonic 92dB third and all other harmonics below 100dB Distortion at 25Hz was slight ly worse with third harmonic measur At 2 83V frequency response was flat limits this can serve as a good line amplifier in other installations But really it needs to stretch out and swing some volts It has anced input showed that the preamp s sonics were of the Burning Amp Festival Photo ing to music Nonetheless when the preamp was plugged into the balanced F4s driving a pair of Basszillas the rush 0 0 5dB from 5Hz to 48kHz my test Newnes 2003 2 Morgan Jones Building Valve Amplifiers Newnes 2005 Though I have not bench tested with other 6SN7s a quick listening test the appropriate interconnects in and out Note that the pream
82. lt Fun listening 46 audioXpress 2 09 www audioXpress com Bamboo John Kaizan Neptune shakuhachi Arakawa Band First Impression Music K2HD LIM K2HD030 This is one of the many albums from transplanted American shakuhachi master John Kaizan Neptune which have been extremely popular in the Far East Neptune was enchanted with the sound of the simple Japanese end blown bamboo flute and studied it formally in both Hawaii and Japan becoming a master teacher performer and composer as well as developing improvements on the instrument s design and building new instruments himself This album was released 26 years ago and won a Japanese Best Recording of the Year award in 1980 The five tracks are a mixture of jazz funk and blues combined with Japanese folk music influences The second and fourth tracks are the longest at over 12 minutes each and feature some exquisite sounds from the various
83. minimum or maximum sample values which should be the same Ready willing and offering an extensive range of ready to go toroidal transformers to please the ear but won t take you for a ride 0 Avel Lindberg Inc 47 South End Plaza New Milford CT 0677 6 tel 860 355 4711 fax 860 354 8597 sales avellindberg com www avellindberg com audioXpress February 2009 35 Fie Edt Wee lest fe Took Options Hee DSF eSB H om OS e lilo SRS ein 1 COS OOO OOOO i Bite Emphases Olt Gis Din Gist 17 ar Fresch Off Gea Ort Bishi L Doke Baphia OFF Ghiteid Meb Getic L bike Eckes O Din Ort De 0 bd Beeps OF CLR Cd 28 bee Dep OT Drit nt Alikin Ie 22 kh Depa OFF CI ON Oe 23 pdt Beep OFF ORAM ELE OOO H likh Eectaen OFF Witi OEE sD m t Beep OFF Geis Goede Zo e Lt Beeps OFF Ciao Ort Zo 21 kH eps O CIS Mbesa aS e dg Daph OFF Ca Oe Gro J St tive Ce Oe GTS GHz 2 St true Ce Dr OPES OCHO 318 tt Emphases Oe CR a Ee ee 0 1 0 Stir freeen Ce rimm Gee Bette 7 GiH Freen Cen Gites Gedo 0 0 Ho POOH Emphas Ce Dr i Gee De SS Sttz Ephes Ce Ona Mati Deo W SOE Emphas Ce OTe I CO D t Ephesi On ORT Oe Sie iph On i COSI COS A mm ise Emphases On Gi Oe OOD en aie phai Cen Ce Oe ici ap Se Eph Ce CMa Meee ZZ 42 G t Eph Ce f ets DZ WOH Ak H Emphasis Cen Wen feel Bet frech et j PER ee es
84. nsformer and want the absolute best performance the pre RS FIGURE 5 A simple yet high performance CCS TubeCap Optimized Audiophile Resitors e Silver Gold MCap RXF Radial Xtra Flat Capacitor _ sop RXF MUNO OPE oa Featuring the ultimate winding geometry edgewise for e extremely short low loss signal transmission extremely reduced residual resistance ESR remarkable low residual inductivity ESL Polypropylene capacitor foil alu metallized Grouted winding against microphonic effects e Fit In Adaptors now available MCap Supreme Aeren MCap Sup MCap Supreme Silver Oil MCap Supreme Silver Gold High Voltage MKP BE en NEW MResist Supreme Varied Foil Coils amp Air Core Coils Matgen Recommended US Dealer MADISOUND SPEAKER COMPONENTS Internal Wirings www madisound com Exclusive Canadian Distributor AUDIYO INC www audiyo com Exclusive Argentinean Distributor SK NATURAL SOUND www naturalsound com ar See more audio innovations on www mundorf com and subscribe for our newsletter info mundorf com OEM and dealer inquiries invited MSolder Silver Gold MSolder Supreme 7 4 MUNDORF High End Components Made In Germany Since 1985 audioXpress February 2009 9 amp is capable of If you use the Jensen I ve saved you some time wrestling with scopes and square waves a 100k contro
85. nts and so on Sound is very good distortion is low and there is no hum even with my sensitive horn loudspeakers Z 38 audioXpress 2 09 www audioXpress com Book REVIEW Sound FX Reviewed by Ed Simon ore money is spent on tooth paste each year than on profes makes us a fairly small com munity So it is a bit unusual for me to get a new book and read two dozen or so names in the acknowledgment section and personally know only one of them OK so some of the names are family and the book is on the edge of the work I do but still only one common source is rare Alexander U Case has written a book titled Sound FX Unlocking the creative potential of recording studio effects a clever title considering that there are so many complete recording books that picking one area where most people just past nov ice would want to improve their skills will help sell books The book contains much more than the title implies COVERING THE TOPIC The author has an excellent grasp of what audio starts from where it can go and what he wants to do with it In many ways this book shares some important characteristics with the film cartoons of the late 40s It can be enjoyed as a basic work but those with more insight will be amazed at how much more there is Case approaches his presenta tion a bit differently than I would ex pect in an introductory book H
86. or Thank you for your inquiry regarding my article hope you ll enjoy it and have little _ trouble building it Third the input is directly coupled design to me Lack of negative feedback _ probably is because there is not enough gain in a single 6SL7 stage to permit it Of course maybe the author thinks _ negative feedback is bad Ron Anderson Hendersonville N C ANY EELA ege deet wer amplifier ZA e sound Hypex UcD amplifier nodules are designed from the ground upsfor true fidelity The fully discrete eircuit affords total control over every detail of operation Add to this custom built parts to make it all happen and a team of committed audiophiles to insure the amplifier delivers on its sonic promises absolute neutrality and transparency while retaining the full emotional impact of the music The amplifier operates in Class D needs virtually no cooling and has the lowest EMlintheindustry Listen to one of the many great audiophile and professional products already using Hypex UcD modules Highlights E Flat fully load independent response E Low output impedance E Very low frequency independent THD E Very low noise E Fully passive loop control E Consistent top performer in listening trials Plus Update any audio circuit with HxR modules ultraquiet ultrafast discrete voltage regulators OEM products available h y p ex electronics UcD70
87. or 8mA of plate current and 160V on the plate the impedance A 6SN7 in grounded cathode mode with an 8mA constant current source plate load will show less than 0 04 THD at 20V peak selected samples _ will show even less This distortion is overwhelmingly second harmonic At 2 83V RMS out the THD is better ing a low 10R AC impedance With than 0 01 again predominantly sec open loop Biasing in common cathode voltage amplifiers typically involves a series re sistor in the cathode circuit with the cathode bypassed to ground A high quality bypass cap is bulky and expen alternative is the use of forward biased diodes for developing a constant voltage at the cathode LEDs are particularly suitable because of their low source im forward voltage drop A cheap surplus 10mA current And it has the side ben Looking at the characteristic curves grid must be about 3 5V with respect to the cathode Or the cathode can be 3 5V positive with respect to the grid So two red 1 7V LEDs in series between cath ode and ground will automatically pro vide about the right bias while maintain constant current loading of the tube the Distributor Wanted D42 1 DIY Kits 2 way design Message in aA BOWS KE eem W passive radiators Web w Wil tb speaker com Wino tb speaker com d audioXpress February 2009 7 the LEDs impedance with current is suppressed Anoth
88. ountry summertime voltages are near the lower end of that 20 Fes 1 2 a R7 CJC This means that I must design the power supply for a transformer that s re ally putting out not 18V AC but 16 2V AC This runs through the full wave 2 I make the deliberately pessimistic as sumption that I ll drop 2 5V in the di odes and of course Pll drop some volt tolerance diode bridge and into a filter capacitor age Vdrop1 in R4 12 Vdrop1 R4 Itotal 0 257V The peak voltage produced by a sine wave is Vac sqrt 2 so taking the losses and low line voltage into ac count 13 Vpk lo Vae lo sqrt 2 25 Vdrop1 16 2 1 414 2 5 0 257 20 15V We ve specified that C4 1000uF and C7 is 100uF so the total input section capacitance is 1100uF give or take a bit How much ripple will there be This will depend on several factors but a good rule of thumb calculation when C is ex pressed in UF and I in amperes is 14 Vripplel 2400 Itotal C 0 561V AC A respectably low number How to com pute the DC voltage Another rule of thumb calculation again with capacitance in UF and current in amps 15 V DC Vpk 1 4170 Itotal Vpk C For this design at 10 wall voltage e that works out to 16 V DC lo 20 15 1 4170 0 257 20 15 1100 19 18V DC I said earlier that the minimum voltage the regul
89. p run single ended 3 Michael Maida High Voltage Adjustable Power Supplies National Semiconductor Linear Brief LB47 1980 not degraded by the substitution of a 5692 or a button base 6SN7GT The test bench performance was all 4 Nelson Pass First Watt Model F4 Opera tion and Service Manual 2006 Ever in a rush to get things done the final wiring took place in the back room 0 04 at 20V peak 40V peak peak out 5 Stuart Yaniger The Heretical Preampli fier diy Magazine 2005 available at www basaudio net 3 This occasioned derisive hoots from passersby and curious stares from those who wondered why we d be solder ing instead of doing sensible things like drinking socializing and listen ing 89dB Hum and noise were below 0 5mV nearly 100dB from full output 6 Stuart Yaniger The Red Light District diy Magazine 2006 7 Joe Curcio ST 70 With Solid State Regu lation Glass Audio Vol 1 number 1 1989 ACKNOWLEDGMENTS As usual the gang at diyAudio com were a great inspiration and help and stress all were worth it dead silent background and no artificial sweeten If the 26dB of gain is not an issue Mark Cronander really got the proj ect started and Chris Bridge did far ing or compression It did everything sonically that you could want from a preamp driving the Pass amps without _ breaking
90. r dware as shown in il me root TaN i DComp 1 498msec A III 24bit 192kHz soundcard d Je abot tke Recto ai a a ia At rota Che e bach eckickk him t have a ER HEI Tae neta iea i ah cite eset leona Setthis fon the power amplifer j Fal l RSA sore MT UE ex Latest version always available ey i aan EECHER Nea fements Quasi amp nechoic Frequer E online Friendly small company EENEI No endless voicemail Bass Restored Response Liberty Instruments Inc carolst one net tel fax 513 755 0252 Visit our website for additional information downloadable demos and freeware www libinst com audioXpress February 2009 11 with the upper leg bypassed by C2 The voltage divider resistors slightly load the voltage amplifier but at 1M minimum the loading is pretty insignificant more than a hundred times higher than the 6SN7 s plate resistance The choice of output coupling caps C3 and C4 will depend on the expected load their value sets the bass rolloff cor ner with a 3dB down point calculated from f3 1 6 28RC where R is the expected load resistance For a 100k load a very moderate 0u47 capacitor will give an f3 of about 3 5Hz Although the F4 s guishes as the preamp warms up and the bas LEDs start to light safely be increased to 100k without wor POWERHOUSE ries about offset voltage or RF pickup The important thing is to have
91. r noise screening At this point Id say that the LSK389B aX Electronic Crossovers Tube Solid State Passive Crossovers Line level Speaker level Custom Solutions We can customize our crossovers to your specific needs We can add notch filters baffle step compensation etc All available as kit Free Catalog Marchand Electronics Inc PO Box 18099 Rochester NY 14618 Phone 585 423 0462 FAX 585 423 9375 info marchandelec com www marchandelec com audioXpress February 2009 27 GEHEIT By Paul J Stamler Tri Way Low Voltage Supply Pt 2 To show the versatility of this supply the author puts the Tri Way board through Its paces PHOTO 1 The Tri Way board with most of the parts installed for use in a compressor vm going to go through three very different designs using the Tri Way UU do the first one the long way showing all the math PI 2 Vadj whiz past the other two stopping only to point out how they differ from the first To keep the equations straight I ve numbered them Before I start I must remind readers that a good power supply design should not only work under normal conditions but also should be built to keep working under worst case conditions These might _ include low or high line voltages compo 4 P Iref R13 nents at the limits of their specified toler ances and high operating temperatures Ever work a festival in 100 heat with
92. rated the audibility of phase distortion at Boston Audio Society and has designed the Omni Focus speaker bipolar coincidental with phase linear first order crossover ARP 2600 analog music synthesizer 1kW biamp and PWM supply at A D S and Class D amps Paul J Stamler Tri Way Low Voltage Supply Pt 2 p 28 is a recording engineer producer musician and technical writer he also hosts a radio program No Time to Tarry Here featuring traditional folk music and related stuff He has delighted in 8s since he was a boy when they were still being made Gary Galo A De Emphasis Test CD p 32 is Audio Engineer at The Crane School of Music SUNY Potsdam where he also teaches courses in music literature A contributor to AAC since 1982 he has authored over 230 articles and reviews On audio technology music and recordings He has been the Sound Recording Reviews Editor of the AASC Journal Association for Recorded Sound Collections since 1995 was co chair of the ARSC Technical Committee from 1996 to 2004 and has given numerous presentations at ARSC conferences www arsc audio org Mr Galo is also a frequent book reviewer tor Votes Quarterly Journal of the Music Library Association has written tor the Newsletter of the Wilhelm Furtwangler Society of America and is the author of the Loudspeaker entry in Je Encyclopedia of Recorded Sound in the United States 18 edition Arto Terho Showcase 20W Push Pull Amp
93. re emphasis specifi cation applied a high frequency boost ahead of the analog to digital convert in playback usually with an analog fil ter after digital to analog conversion Because low level high frequencies remained boosted during the D to A conversion process linearity problems in those converters were also reduced By the early 1990s many manufactur ers of digital conversion chips were im plementing de emphasis in the digital domain usually in the playback digital filters At that point low level linear ity of D to A converter chips had im 32 audioXpress 2 09 proved to the point where it really was under development Sony phasized during the conversion process not necessary to keep the signal pre em The Red Book pre emphasis de em is a simplistic and incomplete explana record and complementary cut in play back does reduce quantization noise but this was probably not the greatest sonic benefit The greatest benefit was the with a resolution of 14 bits had very _ low signal levels By the early 1990s the ticularly in the high frequencies and improved high frequency linearity at ally implemented the Red Book pre emphasis standard Nearly all of the ers which ensured that low level high frequencies would be recorded in a more linear fashion This high frequen cy boost was applied with an analog equalization circuit because it needed to
94. reamp with balanced outputs is used for all of the voltage gain and to drive the cables to the power amps along with the power amp input impedance The F4s buffer the balanced preamp outputs and provide a balanced signal to the speakers Because a unity gain power 3 stage can be designed to run open loop with quite low distortion the perfor mance of the electronic path rests almost entirely on the balanced output pre and characterized the input and driver stages of a push pull power amplifier fit the ticket perfectly I ll return to this LEFT CHANNEL LOUDSPEAKER 8 OHMS F4 AMPLIFIER XLR OR RCA INPUT CONNECTORS RIGHT CHANNEL BALANCED OUTPUT PREAMP RIGHT CHANNEL LOUDSPEAKER 8 OHMS F4 AMPLIFIER FIGURE 1 Balanced mono connection for the F4 audioXpress February 2009 5 point at the end Given the symmetrical loads and the requirements for balance you have sever al options The first that comes to mind is the classic differential amp Fig 2 With a constant current sink in the tail and load that is equal on both polarities the balance is excellent Likewise the output impedance Zout can be made relatively low with symmetrical loads it is roughly equal to the plate resistance in parallel with the load resistance Power _ paid to the cables and amplifier load this could be a viable option albeit at vantage of potentially lower input ca pacitance
95. rives real drivers things do normal tweeter polarity apart which is the same as being in phase has the notch The system phase shift once Sum 1 414 volts HP 41 volt FIGURE 2 Vector summation of two signals at 90 apart 50 100 200 500 100 200 500 1k Frequency FIGURE 3 Performance of second order all pass crossover www audioxpress com frequency of this notch So with ideal all oass COs the pattern should be down 3dB at the CO frequency _ while for even order COs they should be down 6dB As Mr Mogavero has spotted the third order CO tried Fig 29 in my ar ticle was down more than 6dB at the CO Fig 27 in my article shows the lower CO point is down about 8dB This is because not behave as predicted in the ideal The COs used in my article were devel oped by modeling the acoustic response of the system without regard for using ideal _ AP2 or B3 CO shapes The drivers imped ance variations and their response anoma lies make the acoustic LP and HP responses much different from the expected electrical response of the CO components It is the acoustic CO response that you care about and using ideal electrical CO networks _ will likely not produce what is needed in the way of acoustic CO responses How the acoustic CO responses sum is further modi fied by the drivers physical positions and their relative acoustic origin locations Thus
96. rly calls for a bigger heat this board goes up to 2 5 high which provides significantly lower thermal re think it s time to punt by putting the LM337K come in TO3 packages avail able from Mouser and something like a Wakefield 680 125A heatsink mounted on short spacers would dissipate 8W handily I could mount this heatsink on the outside of the cabinet provided I used a plastic insulating cover on the regulator itself since there is voltage on the case Supply 3 for 48V DC phantom power H A all a Kai deck NEW REPRINT OF THE 2 EDITION Starting with lucid descriptions of the parts involved resistors capacitors inductors tubes and transistors the author then begins joining them together to form circuits offering the reader an accessible approach to linear electronics design Focusing mainly on audio amplifiers the book features two of Linsley Hood s classic designs which have become popular with the DIY crowd Includes excellent coverage of amplifiers oscillators power supplies filters and feedback using clear examples and easily solved equations Reprinted 2006 1998 348pp 7 3 8 x 9 5 8 softbound ISBN 1 882580 49 4 Sh wt 2 Ibs SECOND EDITION THE ART OF LINEAR ELECTRONICS John Linsley Hood 30 audioXpress 2 09 BKAA70 For many years do it yourself audio enthusiasts have asked me What s the best book for learning amplifier design Thank
97. ry inductor includes parasitic ca pacitance between adjacent coil turns _ and between start and finish windings These capacitances result in the paral lel resonances shown in Fig 1 Inductors intended for use in loudspeaker crossover using the basic SPICE model even at typically between 100kHz and 2MHz _ larger values resonate at even lower fre quency At higher frequency as shown in _ Fig 1 other lesser resonances should also _ be anticipated For example the 3 5mH inductors used in my horn loaded speak er crossover resonated at 210kHz but exhibited many smaller high frequency resonances unless you take care to leave sufficient un AN IMPROVED _ INDUCTOR MODEL Using SPICE basic components and with e networks will exhibit a major resonance Impedance simulations using pie supplied inductance Model 10K MV Joen WP FIGURE 2 Simulated impedances of 1 156mH and 250uH inductance same values as measured for Fig 1 using SPICE default models BK 16 Kit Madisound is pleased to offer the BK 16 folded horn kit We have chosen the Fostex FF165K 6 5 full range for use in the BK 16 cabinet The FF165K has a Kenaf fiber cone inverted foam surround and aluminum dust cap The FF165K is run full range with a frequency response out to 15kHz The T90A super tweeter has been added to cover the upper frequencies The T90A is a top mount horn tweeter with an Alnico ma
98. s to Edward T Dell and Audio Amateur Inc this valuable work is once again in print and now I have an easy answer to the What s the best book question Nelson Pass To order call 1 888 924 9465 or order on line at www audioXpress com Old Colony Sound Laboratory PO Box 876 Peterborough NH 03458 0876 USA Toll free 888 924 9465 e E mail custserv audioXpress com www audioxpress com Phone 603 924 9464 e Fax 603 924 9467 www audioXpress com to run a couple of microphones Phan tors I specified the largest 100V units that would fit in the space ripple volt age at the regulator input will be about 20mV which is negligible tom powered mikes draw varying cur rents but the worst case would be dead shorts which would draw 14mA apiece Not that you would care about the reg ulation if the mike is a dead short PI also want to draw another 15mA to pre load the regulator Pll get to that in a bit for a total of 43mA With a voltage this high Pll use a TL783 regulator instead of an LM317 and that changes a few things The adjust terminal currents and its nomi are still the same 1 2V and 1 3V Texas Instruments recommends that the adjust terminal not be bypassed with a capaci tor so I can leave out C24 Setting the reference resistor R18 to 1k the nominal current is 28 Iref Vref R18 Iadjust 1 27 1000 0 000083 0 00135A 1 35mA To get 48V
99. sonic metal oxide resistor that s actually draws 15mA under nominal 29 Vadj Vreg Vref 48 1 27V conditions How much heat will R21 dissipate what I d use Finally the regulator will ue e 1 14W under worst case circumstances so on WINDING UP The Tri Way board can be useful for a variety of projects I ve left open as many options as possible in the hope that readers will find new ways to make it sit up and do tricks previous examples this is a circuit where the current drawn by the load will vary a lot Some microphones such as Neu mann KM 84s pull as little as 0 5mA _ from the phantom supply while oth ers draw as much as 6mA Ideally the TL783 has higher dropout voltages and _ all circumstances even when there s no nal reference voltage is 1 27V instead of 1 25V Oddly the tolerance limits As I said when I wrote up the Gamp supply I no longer fabricate my own cir cuit boards the chemicals are more toxic and messy than I want to deal with and I can at least hope that commercial man ufacturers have better facilities to handle and dispose of the stuff safely You may not agree so I ve made a board layout that s posted on the audioXpress website it s as close to the fabricated layout as my graphics programs will allow Or of course you could buy your _board s from me like the Gamp _ boards the pric
100. structive interference and real world drivers don t do destruc time dispersion issue You want to keep it suggestions are similar to mine In my article titled Tweeter Setback Sept 08 I reduced the delay disper tive interference well I agree with your recommendation to use a digital electronic CO because it signs are probably of more merit than true sion by putting the front of the tweeter behind the front of the woofer Your can give arbitrary delay and a wide CO 3 choice I also agree that good drivers are very important I have a pair of Lowther _ knockoffs that sound delightful in spite article because know of no passive imple article and mine have a common goal to reduce the effects of phase distortion Your article is more pristine in that it approaches zero phase distortion My article acknowledges phase distortion but asks what can be done to reduce it to _ forward to your next article I was very interested in your com 3 ments on the audibility of different crossovers You display unusual for hi fi honesty Your praise of the third order crossover agrees with my assessment In inaudibility my opinion fast crossovers help in re ducing lobing There is one crossover design you did not mention that can result in a linear phase loudspeaker In the supplementary crossover CO one CO say the low pass is subtracted from t
101. t 20dB below the 1kHz 20kHz 10 81 108 1 0 951 Compared to my 2006 order in 11 76 117 6 1 029 which three of the ten units had measured amplified FET noise 20Hz 20kHz 9 78 97 8 even with noise BW extend 20Hz 1kHz 4 63 46 3 2 012 out of spec noise all 20 units ing to 20Hz Because noise is uncorrelated with AC line ripple and any other signal for that matter an AC ripple level 20dB below the measured noise produces an error of only 0 04dB NOISE VERSUS CURRENT CORRELATION GRAPH Figure 3 contains 20 circled points num The horizontal scale is the per FET operating current I and the vertical scale is the 1kHz 20kHz noise volt age density of the FET for each one in the package If a single FET were thus plotted over a range of currents the graph would be a downsloping curve 20Hz 1kHz band because within limits increasing cur rent increases transconductance because the FET s channel resistance becomes lower This reduces thermal noise volt age But in Fig 3 the 20 points are fair ly randomly scattered There is a small amount of lower noise versus higher semiconductor physics effects influence the noise level Notice though sam noise and lowest current while it s vice versa for 16 C1 1000 uF C2 10 9uF C3 470P C4 68pF C5 C6 2N2 C7 100 uF J1 IN J2 OUT LED1 67 1612 LED2 516 1358 a1 LSK389B R1 9K435 R2 1K0484 R
102. t can have practical problems In my systems the 105 Odes wrote the article learned that increasing this distance raises the trouble the grille cloth can cause in terms of adding echoes to the tweeter response spent consider able time rectifying this problem finding ally mandatory think Mr Crawford and agree on the limited but there is no need to drive it to zero Thus minimum phase crossover de linear phase designs Mr Crawford brings up the supplementary crossover believe this approach did not fit the intent of my mentation using a single amplifier per chan Since my article George Augspurger has brought to my attention a minimum phase readers to consider It is amazingly simple being a low O second order design with wide overlap You simply design a second with a second order high pass HP with the same O of 0 5 at half the intended cross eb AC Bs BE LA LA IT 104 I 0 20 ao LIL i ig 60100 200 en 100 200 age LI KL LI 20 80100 200 HH UG CITT 600 Ik 2k 6k 10k 20k FIGURE 2 Performance of three way minimum phase cross over with ideal WMTMW array www audioxpress com over frequency Figure 1 shows the shape of the LP and HP responses and the TABLE 1 GEORGE AUGSPURGER S MINIMUM PHASE SECOND ORDER CROSSOVER APPROACH System Crossover LP Design BP Design HP _ Design system performance with a
103. t thing to calculate is the power the regulator will dissipate worst case This happens when the regulated volt age is at its lowest possible value Vreg lo while the wall voltage is at its hottest We already know Vreg lo 14 3V and Vdrop1 0 257V DC for this circuit if the wall voltage is 10 high the trans _ former s secondary will be putting out V AC hi 19 8V AC So 2 293 Vpk hi V AC hi sqrt 2 2 5 A 1W resistor will work and so will a where things dont stay nominal In the What about the ripple R7 and the ca pacitors which follow it 3400uF more or less form a low pass filter _ Vdrop1 19 8 1 414 2 5 0 257 25 25Vpk 24 V DC hi Vpk hi 1 4170 Ttotal Vpk hi C 25 25 1 4170 0 257 25 25 1100 24 27V DC 25 Vunreg hi V DC hi Vdrop2 23 34V DC So the voltage across the regulator will be 23 34 14 3 9 04V DC And the power dissipated by the reg ulator worst case will be 2 32W McINTOSH MARANTZ HARMAN KARDON WESTERN ELECTRIC TRIAD ACROSOUND FISHER CHICAGO STANCOR DYNACO LANGEVIN PEERLESS FENDER MARSHALL ELECTROSTATIC TRANSFORMERS FAIRCHILD 660 670 RCA LIMITERS WE DESIGN AND BUILD TRANSFORMERS FOR ANY POWER AMPLIFIER TUBE PHONE 414 774 6625 FAX 414 774 4425 AUDIO TRANSFORMERS 185 NORTH 85th STREET WAUWATOSA WI 53226 4601 E mail AUTRAN AOL com NO CATALOG CU
104. the Denon Audio Technical CD 38C39 7147 have sweeps with pre emphasis from 20Hz to 20kHz but these are only useful if you have measurement equipment that can be synchronized with a sweep gen erator ROLLING YOUR OWN I have always been frustrated with my inability to measure the accuracy of playback de emphasis so I decid ed to take matters into my own hands Book pre emphasis curve which I pro so the 50US time constant is produced this value is unimportant I used this simulation model to gen without pre emphasis in other words erate the curve shown in Fig 1 Cir allow you to put cursors on the generat ed graph and measure one level relative to another Table 1 shows the measure ments I made on the simulation in Fig 1 The FREQ column lists the fre quencies I decided were appropriate for a truly useful de emphasis test CD I set the first cursor at exactly 20Hz then moved the second cursor to the re maining frequencies and noted the dif FIGURE 2 Simulation model for generating the Red Book pre emphasis curve The 50uS time constant is the product of R1 R2 C1 15us is equal to R1 C1 Preset Untitled 10 0 dB 31 62 CH Waveform shape Length 0 001 to 600 0 seconds Start Frequency 0 01 to 22 050 Hz 16 000 00 End Frequency 0 01 to 22 050 Hz Log Sweep Amplitude
105. the input stock input resistor is 47k its value can resistors as well matched as possible be more than you need and spending time with an ohmmeter R13 and R14 hold the output voltage at ground in absence of a load Their resistors 1 is good enough NE 1 serves to limit the grid to cath ode voltage of the ECC88 at power up _ This is often accomplished by the use of a diode forward biased at power up but reverse biased after warm up And if a neon bulb is difficult to source you can use a diode with its cathode banded end connected to the cathode of V2 _ The advantage of the neon bulb besides a visual indicator of power up is that its capacitance is lower and more linear made guest appearances in such designs with a nice orange glow when you first Morgan Jones s Crystal Palace and my than a reverse biased diode It lights up turn on the power switch then extin I have endeavored to make this circuit uncritical of the power supply but it is cause they are part of the AC load of the split load inverter As before 0 1 matching is easy to achieve by buying still worthwhile to try to get to the point of diminishing returns The huge power supply rejection afforded by the CCS loading of the input stage and the large _ voltage swings make this a very straight forward exercise You need a stable 350V for the B rail the constant current of matching is less
106. ting Loudspeakers by Joseph D Appolito Tollfree 888 9249465 Phone 603 9249464 Fax 603 924 9467 E mail custserY audioXpress com www audioXpress com audioXpress February 2009 15 r within SPICE these frequency and amplitude dependent pa Howeve using an oscilloscope When modeling in the time domain you can modify SPICE models to account for amplitude nonlin earity You can also perform small signal AC frequency domain simulations and lent simulator you can modify capacitor resistor and inductor models to also ac count for parameter changes with fre when using SPICE or a SPICE equiva quency y equency r just like probing the circuit NDUCTORS CIR nductor ini pedance frequenc inductor imp Most SPICE analyses are made using Ge core ir large signal transient simulation to pro d ORY lb ZO A af models can simply be produced using inductor im only the basic primitive elements found in every SPICE simulator cuit s behavio asure u Inductors ir core in the author introduces a hands on approach to SPICE circuit simulation erri v Output2 I R14 v Output5 I R d suited for use with the small values used when H By Cyril Bateman ing ions Electron eon lat the Simulation Program with Integrated Circuit Empha has become by far the most widely used circuit simulation 250uH Fal 29 PICE SIS v Outp
107. tly twisted and dressed close to the chassis away from any low level signals As is customary the heater supply center tap is DC biased above ground but AC bypassed to ground by C109 This minimizes hum pickup in the to cathode voltage within the specified maximum WE CAN BUILD IT YES WE CAN modular boards with dif ferent construction tech niques The circuit is simple enough to be done point to point or on a pertboard if there is inter est it would be relatively easy to put the whole thing on a PCB Photo 1 shows the signal circuitry built on a perfboard Because of their size I put the output caps C3 and C4 on a sep arate board positioned be tween the main signal cir cuitry board and the output connectors Photo 2 shows the raw supply and regulator built on a mixed platform of perf and PCB As is evident 14 audioXpress 2 09 Signal circuitry built on a perfboard OG s KS ei Ce Lei there s nothing exotic in the componen _ try nor given the pains we took in the mode noise to chassis ground they are optimally placed right at the 6SN7 circuit design is there any need frequencies PR rae Pi B y D See bad P s w a a St i ay Wl Ab P m e Pa e Fa J PHOTO 3 Final wiring in a back room at the Burning Amp Festival and before clean up www audioXpress com e PHOTO 2 Raw supply and regulator more space than I needed Unfort
108. tor R3 in the test circuit the current can be as high as 8mA per FET 16mA package total while providing a drain voltage V of at least 2V The range of FET currents measured here is comfortably within the range that s allowable for proper and low CU EH HE H ie SS pm A i j iT KR i i F gie L i 7 ttt i pe mm pm ON D ES Tt tittle Le SRES 20005008 cote Sesspssd Fees Eeges Hite RSR EEG a i LCE BRRSSSSS m GE BREE REE BEN Za GE KE GELD ECK LUX a an ak fm e Hi E i Ii i z i D i LX H D l io ALE ZP n LE t E A P I H a Los gt re D D afe oo oo i a fee ema faa 5 E kr z i Tetti EE DE a hH f e f d ei Hri E i EEEE a d D di HH a bes 12 13 2007 10 35 TEL 6033644414 D M GT o lt wi E noise operation in the LP797 phono large source bias resistor from a nega tive voltage supply I didn t do this in _ the phono preamp because I didn t want an electrolytic cap large value 1000uF required in the audio path But for the _ ripple by temporarily shorting the Fig 1 test circuit output and then connected noise measurement preamp that s fine noise measurement BW is for all practi _ cal purposes determined by the settings of the low noise measurement preamp following the test circuit To achieve freedom from AC line hum corruption I used a very well fil
109. u nately the inside height is just too small to accommodate the CV1988 and the Grounding is straightforward l used a set of individual stars all brought ro a common point much as indicated on the schematic The common point is connected to the chassis safety ground at one point via a groundbreak ep which helps prevent the formation of ground loops with other equipment _ The groundbreaker is bypassed at RF by C108 whose value is small enough 6SN7 and keeps the ECC88 heater to have negligible conductance at hum also look cool and work flawlessly I found a great source of casework my own weakness at DIY Enclosures _ LLC www diyenclosures com I used For the prototype I used a mix of their DIY1712 cabinet which had way room needed to clear the pins from the bottom of the case This is the reason for the vertical mounting of the PCB I chose the position so that the tubes were under the vent slots in the case s cover I drilled a few holes in the bot tom plate to aid circulation and as a 2 result the preamp runs only moderately warm Output connectors are XLR just to emphasize the balanced nature They ON FURTHER EXAMINATION One nifty feature of the split load in verter is the opportunity to pick off a single ended signal from the cathode So if you want to drive an F4 or other amp single ended you can do so For the output stage to run as a cathode fo
110. uency tand of course is frequency dependent The capacitor electrodes and connection resis tances exhibit inductance which also acts in series with the capacitor A reasonable estimate for self induc tance is to allow 7 5nH for each 1cm length of straight leadout wire plus a slightly lesser amount for each 1cm of the capacitor body length Should the maker s impedance plot be available then you can more accurately calculate self inductance knowing capacitance value and the ca pacitor s self resonant frequency At self resonance the capacitor s reactance and self inductance become of equal value but opposite phase angle so they cancel out l d d Awais Hence L The most nearly perfect capacitor near perfect capacitor must also halve In dependent almost but not quite halv octave increase in frequency At some The simplest fixed value three com Making these parameters frequency de pendent extends the model useful range but negates use for transient simulations _ By adding a second larger capacitor and resistor you can realistically model over a more useful range both for AC and tran sient simulations while retaining fixed a practical capacitor ESR is frequency easily calculated either from makers pub lished graphs and data or from measured _ values Fig 8 Similar schematic models are provided as standard in the better non SPICE based simulators element val
111. ues The required values are In a similar fashion provided the mak ers impedance curves are available you inductors used in crossover networks such data usually does not exist You could measure the resonant fre quency and impedance at resonance using a signal generator and suitable test meter ot lacking this equipment you could even make a vague estimate for resonant fre quency based on the inductor value and the inductor plots shown Despite not producing an accurate model even that Test Equipment Depot j e F 9 Cp 800 996 3837 The source for premium electronic testing equipment New amp Pre Owned e Oscilloscopes e Power Supplies e Spectrum Analyzers e Signal Generators e Digital Multimeters e Network Analyzers e Function Generators e Impedance Analyzers e Frequency Counters e Audio Analyzers Call now to speak to one of our knowledgable salespeople or browse our extensive online catalog www testequipmentdepot com e Sales Repair amp Calibration Rental ev Leases e Buy Surplus audioXpress February 2009 21 INDUCTOR CIR Impetiance of inductor model schematic cs bd Complex madels Blue lines simple models Red lines les ll ell ol 10m 5m LIDL nt zt LIDL aa 10K 00K 1M V V_cap_simple I R71 V V_cap_with Per V V_cap_Maplin I R73 FIGURE 8 Using the schematic models in Fig 1 you can quickly produ
112. ut1 VIS v Output4 I R22 o Bn D gt O o o yo Bn O Bn D X o Q N Oo 5 Le o o 5 Bn O tj Oo D E D o E zo D Bn O O o se Bn D Ke Oo Z e Oo D Bn G re O D e c ES O a o Bn ro J O 5 Oo E o o e Ke Ge O o O Z o Q E Oo D H 5 o O Lu DE m D 0 00K G and Networks program It is provided with competent transistor macromodels so performs well when simulating integrated circuits Basic capacitor resistor and inductor models are also included but only as idealized Using the Micro cap software theoretically perfect components S S to devise new improved user models able to accurately mimic inductor behavior by frequency E si KC EN N 3 N www audioXpress com audioXpress 2 09 16 rameters are mutually exclusive you can not use both parameter sets within one simulation Of course SPICE is not the only type of simulator other simulators model prin cipally in the frequency domain and many can combine both time and frequency do mains together by using convolution as in the Harmonic Balance simulators Such simulators are usually provided with a li brary of truly competent inductor models able to replicate real world components even to very high frequencies INDUCTOR MODELS Unfortunately few component makers provide realistic SPICE mo
113. ws _ the system output for both drivers wired the same polarity solid line and for the tweeter HP polarity inverted dashed curve The fourth graph is the system phase shift for the same conditions All plots are for a 1kHz CO frequency The B1 CO 6dB octave is shown in Low Noise lt 1nV Monolithic Dual Tighter ldss Matching Narrow Idee Grades Low Capacitance 20pf QUALITY PCB amp SERVICE PROTOTYPE TO PRODUCTION instant online quote shopping cart ordering system China competitive prices free electrically test http www pcbcart com sales pcbcart com 86 57 1 87013819 86 57 1 87036705 No 76 GuCui Road Hangzhou China CS ze WWW PCBCART COM eee Sa NRN 100hz Ju Frequency Wkhz 100khz Functional Replacement for 25K389 wWW linearsystems com 000 339 4025 audioXpress February 2009 41 Fig 1 As Mr Mogavero notes this type CO has both responses down 3dB at the CO frequency Note that the phase dif ference between the LP and HP is 90 throughout As Fig 2 shows when two equal sig nals are vector added with 90 spacing the result is 1 414 times each signal or up 3dB This is why this CO has the LP and HP down 3dB at the CO point Note that with either tweeter polarity the system sums to unity However only with normal tweeter Thus the normal tweeter polarity sums to unity while the inverted tweeter polarity
114. y side I used a power entry module for the fuse switching and IEC line cord plug You can use individual bits for this too Whichever earth ground lead from the power line is firmly connected to the chassis Safety is the number one concern Of the mil lions of ways of making a raw supply this is certainly one of them As for heaters the ideal solution is a separate heater transformer so that rec Dier noise from the high voltage sup ply isnt coupled to the heaters I took a for DC here the signal levels are high and with some care in layout and lead The SP495 Audio Consultant will help you confidently install and analyze all your sound jobs Call 1 800 SEZNICORE sy ugu N48v 40 0 40 OVN s Les A wed RT6O Oms SOUNA Audio Consultant Quality Stability e Total Business Solutions SENCORE 3200 Sencore Drive Sioux Falls SD 57107 736 2673 Walk into a venue with the SP495 Audio Consultant and instantly begin making sound measurements quickly make system adjustments and confirm proper operation all with complete confidence Visit our website www sencore com sales sencore com 1 800 736 2673 or 1 605 339 0100 audioXpress February 2009 13 Wii M cl l H Wy FFFFFEFFEEH G PHOTO 1 dress you can avoid hum Capacitors C106 and C107 bypass any common heater pins though I sited them clos er to the transformer All heater leads should be tigh

showcasing a 20w push-pull amp cd to measure de

Contents

Download Pdf Manuals

Related Search

Related Contents