AudioNote-SP12Ankoru-pwr-sch维修电路原理图.pdf

Issue 12 - SOUND PRACTICES 9Hey, dont drop thatthing on your foot!Musical information is a dynamic fourdimensional continuum, like the one posedby relativity theory, consisting of three spa-tial dimensions and time, all inseparablyinterrelated. A system for music recordingand reproduction must transfer this con-tinuum and faithfully reconstruct the orig-inal sonic performance in the listeningroom. Test instruments are supposed toensure that the transformation of the con-tinuum is linear and accurate to certainparameters.Alas, these instruments and the mathemat-ical models that we use, such as relativitytheory, lack the spontaneity and emotionalcontent vital to music. Somehow ourprimitive method of recording scratchesinto a vinyl disc captures some of thisemotion, and the lump of rock we call astylus is able to extract the informationand convert the vibes into a signal readyfor the amplification chain. The amplifier,therefore, must not only perform wellelectrically. It must also convey emotion inorder to fully satisfy both the analyticalmind and the inspirational soul.These days, science is beginning to discov-er an essential, almost mystical, intercon-nectedness of everything. It is intuitivelyobvious that the character of the universeon a macroscopic scale should rely on theproperties of the subatomic particles ofwhich it consists. At the same time, thecharacter and properties of those particlesis defined by the universe at large, thewhole system mysteriously holding itselfup by its own bootstraps, each piece ofthe giant jigsaw fitting exactly into placewithout deficiencies or excesses.It is only by virtue of an intellectual ges-ture that we perceive a condensed, solid,and definable part of the web of reality,yet we have deceived ourselves into think-ing that our mental creation is the be-alland end-all of existence.Most of our old scientific “laws” includingthose currently used to judge sonic perfor-mance are only close ups of the wholepicture. Im afraid we are not seeing thewood for the trees.Certainly these measured parameters dohave some relevance in terms of overallperformance, but to recreate a musicalevent, an amplifier must work on both amacroscopic scale as a part of a communi-cation system between the performanceand listener and on a microscopic level as aAnkoruby ANDYGROVEAudio Note, U.K.RadioFans.CN 收音机爱 好者资料库10SOUND PRACTICES - Issue 12collection of valves and parts which mustbe tamed and optimized for the task athand.A magazine article can only skim the sur-face of any design philosophy and, ofcourse, there will be shouts of “what the*$! is this guy on?”, but I hope my dis-cussion of the Ankoru design will be inter-esting nonetheless.Starting with the basic precept that eachpart of the amplifier should fit exactly intoplace, and have a character defined by theoverall requirements of the system, thevalidity of feedback and push pull opera-tion, two pillars of traditional amplifierdesign, are immediately called into ques-tion.These concepts are purely intellectualconstructions, created in laboratorieswithno motivation from natural music, and Iam convinced that they detract fromsound quality as a result.In practice, the ultimate purpose of feed-back and push pull operation is to makeamplifiers easier to make not better. In anyevent, reducing harmonic distortion tovanishingly small levels and increasingbandwidth from DC to cosmic rays doesnot make a more musically satisfyingamplifier. Specs must give some satisfac-tion though, cos we all know a guy whoslinks off to the bathroom with a copy ofhis tranny amp spec sheet!I agree that limitations such as distortionand bandwidth abberations unquestion-ably colour the sound and should be elimi-nated, but beyond that I maintain thatthere are more important areas to be con-sidered if musicality is the ultimate goal.According to my way of thinking, all of theabove leads to the assertion that the over-all topology of an amplifier must be single-ended and there must be no feedback.Transistors and all things silicon soundunnaturalput sand in the signal path andyou get gritty sound! So, lets proceeddirectly to valves and, in particular, thesimplest and purest amplifying device.The materials used for the construction ofthe passive elements of the amplifier arejust as important since the signal must passthrough them. Every material has a tonalcoloring effect, so only highly-specified,high-purity, listening-tested materials aresuitable.For example, in the Ankoru we use onlyBlack Gate and Cerafine electrolyticcapacitors for the audio circuitry. Thesecaps eliminate that electrolytic mushinesswithout going over to the brashness of cer-tain plastic caps. The valve selection wasguided by the notion that the differentsonic signatures of each type should becomplimentary, leading to the goal of asound that possesses both strength andfinesse.Schematic diagram of authors design - Audio Note Ankoru SE parallel 845 70 WatterRadioFans.CN 收音机爱 好者资料库Before I go on to describe the Ankoru cir-cuit in detail, I would like to say a fewwords about transformers and transformercoupling, since transformers play animportant role in the design.In any valve, waveform distortion is causedby the characteristic parameters of thevalve changing in sympathy with theapplied signal. In a standard RC coupledtriode circuit, the valve is set up with aquiescent current (Iq) flowing through itand the load resistor, yielding a particularquiescent voltage on its anode (Vq).With a negative-going input signal, thecurrent is reduced and the anode swingspositive due to the reduced voltage dropacross the load resistor (Rl X Iq). Thereverse is true with a positive-going inputsignal, the valves anode current isincreased so the voltage on it reduces dueto increased drop across Rl.There is a problem with this, however,because as the anode swings positive andthe current decreases, the transconduc-tance of a valve goes down due to the cur-vature of its characterisic. Of course, thereverse is true with a positive-going inputsignal, the transconductance goes up withwith the current.This means that the positive part of theanode swing is compressed and the nega-tive part is expanded waveform distor-tion. Usually, this distortion only becomesserious with very non-linear valves and/orlarge voltage swings. When we want todrive a fairly meaty output valve we needto swing a lot of volts because the mu ofthese types is necessarily low to keep theanode impedance down to keep loud-speaker damping up. In this circumstance,waveform distortion can easily rear itsrather ugly head.We need a system for keeping the currentthrough the valve as constant as possibleover the anode swing, i.e., a high loadimpedance. Increasing the load resistor onan RC coupled stage can only go so far,however. One soon runs into problems ofresistor dissipation and PSU voltage if theanode current is kept at the optimumlevel.The SRPP stage and his other activeloaded cousins, such as the mu follower,have never really delivered the goods forme. Close listening reveals a lack of focusand immediacy compared to even thehumble RC coupled stage. Anyway SRPPis a feedback device and quite often thatscheme doesnt work very well electricallyeither, especially with the low impedancevalves we would like to use as drivers.Simply pretending that youve got a lowoutput impedance just doesnt cut any icein the world of real audio.For large power valves, a low AC driveimpedance is necessary because largevalves have large and therefore highlycapacitive grids. Thankfully, the low gainkeeps down the Miller Effect, but its stillthere, so for good HF response, there is nogetting around using a good low imped-ance driver.From the standpoint of sound quality, for astrong sound we need a beefy, low imped-ance driver. Wimpy driver equals wimpysound. Drive two 845s with an ECC83and itll be like putting a model aircraftengine in a Chevy Impala. Not exactly aweinspiring.The DC resistance of the grid circuit mustalso be kept low to control the effect ofanother rather annoying bugbear, grid cur-rent. Unfortunately, the vacuum in manymodern valves is far from perfect, so thereare quite a few gas ions floating aroundinside the bottle. Some of these ions willcollide with the grid and draw electronsfrom the grid circuit. If the grid resistanceis high, the grid bias will be modulated intune with the signal, a real no-no in mybook.Also the grid may occasionally be drivenpositive on signal peaks, causing the grid-cathode diode to conduct, rectifying theinput voltage in the manner of a shuntdiode supply with the coupling cap as thereservoir. This action makes the bias volt-age more negative, reducing the quiescentcurrent through the tube, sometimes tothe point where it will only conduct onpeaks (Class C). In fact, a severe peak cancause the amp to cut off altogether, result-ing in a total loss of output.Worse still, the grid resistor/coupling capcombination acts as an RC time constant,so the effect lasts for sometime after theoverloadhaspassedin sortofa time-delay distortion mechanism.Reducing the grid resistor to combat theseeffects is no solution. We want a DC gridresistance similar in magnitude to theimpedance of the driver valve, i.e. a fewhundred ohms not a few hundred kilohms.Making your grid resistor 600 ohms willlikely kill the driver stage and, anyway,would require a coupling capacitor so bigthat the RC time constant would put usright back where we started.To cure the voltage swing problemrequires a circuit element which has lowDC drop but a high AC impedance. Plus,we need a low DC resistance in the grid ofthe output valve. And the device shouldefficiently couple the driver valve to theoutput tubes grid.The driver transformer is exactly what weneed for the job. Its primary inductancepresents an extremely high AC impedanceto the driver valve and reflects the anodeimpedance of the driver into the grid cir-cuit of the output valve. A good drivertrans will have a primary and secondaryDC resistance on the order of 300 ohms,so the problems associated with grid cur-rent are more or less eliminated. This is aresistance 1000 times lower than Ive seenin some designs.Ideally, the transformer secondary is leftunloaded, i.e. there is no “damping resis-tor” put across it to cut ringing. Anunloaded transformer sounds better and itgives the driver valve a higher impedanceload.There are two large-scale problems withdriver transformers,HF frequencyresponse and LF frequency response.Thesetwo requirements are mutually exclusiveto a certain degree and many commerciallyavailable transformers sacrifice one for theother. The Tango transformers, for exam-ple, seem to go for impressive-looking HFspecs but they have diminutive primaryinductancesIssue 12 - SOUND PRACTICES 11which limit the LF performance.The problem is compounded by the unbal-anced DC current imposed by SE opera-tion, which requires that the number ofprimary turns must be increased to coun-terbalance the loss of permeability causedby the air gap in the core. Leakage induc-tance is proportional to the square of theprimary turns so its a real pain in the butt.The driver transformer in the Ankoru hasto handle 45 mA and still have superbbass, so it took some heavy calculator workand a few trees worth of paper to get it allworking! The Ankoru interstage trans willbe available as a DIY parted.I love the sound of large triodes like the211 and 845. The 845 was used in thisamp because it offers greater power inClass A1. The 211 is a more voltage sensi-tive valve than the 845, its mu is higherbut then so is its internal impedance. Itcant swing a lot of current at the low volt-age end of the anode swing without havingthe grid driven positive into Class A2.When pushing the grid above zero volts, itno longer reacts as a high-impedance ter-minal. It starts to draw appreciable cur-rent, corrupting the input signal in a mostunattractive way unless the driver imped-ance is extremely low.The grid-cathode diode impedance of a211 is about 2k, so we would need some-thing around 100 times lower or hideousdistortion would result. The waveform dis-tortion could be corrected using feedbackbut why build an amplifier that is intrinsi-cally nonlinear?The 845 can sink a lot more juice wherethe 211 starts wheezing, but since the muis so low, it requires a driver stage capableof considerable voltage swing. The 845s inthe Ankoru are biased at -100 to -200Volts for an anode current of 75 mA at1200V B+, they look into a load imped-ance of around 6k, and put out a formida-ble 70 Watts. The output transformer hasto cope with 150 mA DC and hold its 6kimpedance down at LF, requiring a highprimary inductance. This takes a serioushunk of iron, but the Ankoru output is justsuch a beast and the bass is awesome, if Ido say so myself.To keep the drive signal to the outputvalves clean requires a driver valve ofexcellent linearity. One could use an indi-rectly heated valve such as the 6BX7, verylinear, or the slightly less linear 6BL7, butlow impedance, low mu directly heatedvalves are definitely the best choice.Since this amp has to be built using valveswhich will be available for some time intothe future, so that replacements can bemade throughout its life, it was necessaryto use modern versions of either the 2A3or 300B. I originally experimented withthe 2A3 as I wanted a measure of its clari-ty and immediacy, but these valves have avery nasty habit of making toilet relatednoises even in the output stages of amps,and using one as a driver was impossible. Ieven tried some NOS samples but manywere only marginally better, only the bestand therefore rare and expensive exampleswere quiet.So the 300B was chosen, and it brought itscharacteristic warmth and musicality tothe amp as well a greater impact to thebass. The 300B is operated with 300Vacross it and an anode current of 45 mA soit will last for ages, no more current orvoltage was necessary for driving the 845sto full output. The 300Bs output is in factso large that the 845s will be freaking (andso will your wife and the neighbours)before it runs into trouble which makes itsjobs and the job of the input stage easier.Various input configurations were tried, allusing the E182CC/7044 valve for its pow-erful sound. The original and best sound-ing configuration gave the amp so muchgain as to be impractical. Long speakerleads acted like antennae and transmittedthe amps output into the input leadscausing instability. Super high qualitycables and careful system setup wouldeliminate the problem but as this is a com-mercial amp it has to be dealer proof so asimple, single stage RC coupled affair wassettled upon.The 7044 was always run at ahigh current to really bring out its flavour.The Ankoru is interfaced to the preampvia a coupling transformer to allow bal-anced operation and to properly groundthe grid of the red hot 7044. A switchingsystem permits regular unbalanced inputas well. The Ankoru is intended for usewith the Audio Note M3 which has out-put tran