The Pearl II Phono Stage By Wayne Colburn Introduction Here is the long awaited sequel to the Pearl phono
stage, named after my maternal Grandmother who was good with a sling shot , played piano and organ and occasionally listened to my early Hi-fi efforts if the smoke wasn t too bad! Pearl "" addresses the ma#or request recei$ed o$er the years - more gain for mo$ing coil cartridges! Phono stage design is difficult to do well! %$erything matters! &ny low le$el phenomenon occurring in the parts of a phono stage shows up greatly amplified at the loudspeaker! 'his phono stage has about (( dB gain at ) *h+, and another ,- dB at ,- H+, and when you add )- dB or so on a line stage preamp and another .- dB with a power amp you end up with ))( dB of $oltage gain, which is getting up toward a million to )! 'he problems are different than for power amplifiers / problems due to resistor noise, semiconductor noise, capacitor noise, contact noise, stray fields, little bits of hysteresis, radio frequencies, all competing with the audio signal in greater proportion! Circuit topology, choice of components, board layout, regulation, and shielding become more crucial to preser$ing audio quality! The Circuit 0igure ) shows the simplified schematic of the Pearl ""!
'he input signal from the cartridge appears at the Gate of low noise 1fet 23, which produces a $ariation in current through 23! 'his current $ariation is fed to the emitter of transistor 2. which operates in Common-Base mode, forming a cascode connection for 23! Cascode transistor 2. does not significantly contribute its own characteristic, nor does it add much to the gain / it is used to shield the 4rain of 23 from the output $oltage consequences of the input signal, lowering distortion and impro$ing high frequency bandwidth! 'he operating point 5bias6 of 23 is self setting by the depletion mode character of the 1fet, and 2. biased by a constant low noise 4C $oltage deri$ed from the 4C power supply! 'he resulting gain is appro7imately .( dB! 'he output $oltage for this initial stage appears at the collector of 2., where it dri$es a passi$e 8"&& equali+ation network! 'he result of this equali+ation appears at the Gate of 29, the input 1fet of the second gain stage! 'his stage has a flat gain response of 9- dB! 4ifferential input pair 29 and 2( are biased by a constant current source from the negati$e supply rail, and they dri$e the Gate of P channel :osfet 2,, also biased by a constant current source! ;ike the first Pearl, the input stage in$erts the phase of the signal! <nlike the first Pearl, the second stage is non-in$erting, with the result that the output of the phono stage in$erts phase! "f you care about absolute phase, some accommodation must be made in the system to see to it that the absolute phase at the loudspeakers is correct! "f none of your other circuits in$erts, then you would need to either re$erse the speaker leads or the cartridge leads! 0igure , shows the complete schematic!
Here you see the additional details that actually make it work! 'he raw power supply $oltages are regulated by <) and <, and their associated circuitry! 'he unregulated ,3 to .( $olt rails are 8C filtered, acti$ely regulated down to ,9 $olts, and then passi$ely filtered again before reaching the rails of the second gain stage! 'he first 5input6 gain stage is additionally filtered acti$ely by 2) and passi$ely by C3 to remo$e that last tiny bit of noise! 8,- is a user pro$ided resistor $alue that sets the input load for the cartridge if not the default 9= *ohm! 23 through > are matched parallel 1fets making up the input stage! Paralleling these makes for higher bias current with lower distortion and also lowers the noise by about 3 dB o$er that of a single de$ice! 'he 8"&& equali+ation circuit also uses parallel capacitors to get the desired precision $alues! 'he full schematic shows the details of the constant current sources of the second gain stage which uses a red ;%4 to pro$ide the )!= $olt 4C reference dri$ing 2)- and 2))! 'his means that the current will be set by the $alues of 8,3 5)? @ 9>> Ahm B , m&6 and 8,> 5)? @ )-ohm B )- m&6! C= and C> pro$ide frequency compensation! Potentiometer P) is used to ad#ust the 4C offset at the output on 8)=! 'he o$erall gain of the circuit as shown is (( dB C )* H+! 4ecreasing 8)9 from )* to .-Ahms gi$es )- dB more gain when desired for lower output cartridges! 8)( which normally would be shorted to - ohms can be replaced by a resistor to pro$ide more gain as an alternati$e to reducing the $alue of 8)9! C)( has been placed on the board to allow for additional ad#ustment of equali+ation! "n both cases these $alues are up to you! Measured Performance 0igure . shows the response de$iation from ideal 8"&& specification!
'he cur$e comes in about plus or minus an eighth of a decibel from (- H+ to ,- *h+, and reflects my decision to pro$ide subsonic rolloff! Dou can increase the $alue of C), if you would like more bottom end!
0igure 9 shows the distortion at ) *h+ $ersus input $oltageE
Below about ) m? input, the distortion continues to decline monotonically, and the cur$e begins showing the relati$e proportion of noise below that le$el! 0igure ( shows the ,nd harmonic character of the distortion characteristic, reflecting the single-ended Class & nature of the circuit! 'his picture was taken at ) *h+ with an input of . m?!
Construction 'o make it easier for you to build this preamp, we are offering a pc board plus low noise matched 1fets at www!passdiy!com! We supply these items because they would otherwise be difficult to get! 'he board is unique, and the 1fets ha$e been discontinued by 'oshiba! %$erything else you should be able to obtain through distributors such as 4igikey and :ouser! 'here are of course other suppliers, and in the case of the 1fets you can substitute other parts, including ,F*)=- and the ;inear 4e$ices ;F*)=- replacements! %$eryone seems to ha$e their own fa$orite resistors and capacitors! 'here is nothing particularly magical about these parts / they #ust work well and are commonly a$ailable! Dou should feel free to play with whate$er special components appeal to you! " used mostly 8G((4 4ale 5?ishay6 resistors and a number of Wima polypropylene capacitors, and you will ha$e no difficulty finding more e7otic and e7pensi$e types! 0igure 3 shows the component placement for the circuit board! Please note that this pc board is for one channel!
0igure = shows an e7ample of the power supply circuit! 'his is by no means the only way to do it, so feel free to substitute any reasonable supply that deli$ers the required $oltages 5,H to .9 $olts at )-- m&6! Dou may also elect to use switching supplies, wall-warts, stacked batteries / whate$er safely does the #ob! 'his supply is designed to work with both channels!
Here we see the usual components found in a 4"D linear power supply, with a couple items which directly address the needs of this pro#ect! 0irst, there are two rectifier bridges instead of the usual one! We do this because the positi$e rail will be sourcing more current to the circuit than the negati$e, and in this circumstance dual rectifiers will a$oid the transformer noise which often accompanies this! &lso we see that the chassis is directly 5and with hea$y wire6 attached to the earth ground of the &C line and that the signal ground of the circuit is attached to chassis through a high current bridge rectifier as shown! 'he bridge is for safety purposes, and we use a standard .( amp type to ensure that the &C line fuse breaks before the bridge! 'his a$oids ground loops with the rest of the system sharing an &C line ground, while pro$iding an electrical shunt abo$e != $olts in e$ent of isolation failure! "f you find that it lowers the noise, feel free to also connect chassis to circuit ground with a resistor, capacitor, or wire - but lea$e the bridge in for safety! .(I $olt capacitors are recommended for C, and C.! C) is for noise suppression and since it is directly on the &C line it should be specifically line rated! 'he turntable ground should attach to the signal ground, not the chassis ground 5unless you find that it is quieter that way6! Dou probably want to enclose the preamp in a metal chassis to shield it from stray magnetic and radio frequencies! 'he power supply circuit should be at some distance from the phono stage circuitry, otherwise noise will be picked up! 'he best approach is usually to put the power supply in a separate enclosure, and if you do this, then %arth ground the power supply enclosure and run #ust the three wires to the circuit / plus, minus, and circuit ground, grounding the gain stage chassis to signal ground! 4on t set the gain stage chassis near the power supply!
Here is the partial Bill of :aterials for one channel of the components found on the pc board! 8esistors are -!,( watt
Conclusion Fo there we ha$e it! " hope it was worth the wait! WayneCpasslabs!com J ,-)- Wayne Colburn!
