of 12

Please download to get full document.

View again

All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
12 pages
0 downs
EUVL I/V circuit
  Zen  Cen  Sen- EvolutionofaMinimalisticIVConversionCircuit EUVL Howitstarted IhavetoadmitthattheAD844basedRIAApreampcircuitfromLCAudio [1] fascinatesme. Iwaslook-ing, once again, at their“optical supply”MC head amp [2] the other night, only to find out that it wasadopted from Leach’s common-base MC head amp [3] published a while ago. TherearealreadydetaileddescriptionsofhowthecircuitworksatboththeLCAudioandtheLeachwebsites, so I am not repeating that here. What interests me about the Leach head amp is that it isnot a voltage amplifier. Rather it is a current conveyor, at the output of which the current is con-verted to voltage by means of the output resistor (Riv). There are a quite few threads on DIY Audioabout the Leach Head Amp, and various people have commented about its noise level and distor-tion level, not always positively.Itstillappealedtomesomehow,becauseofitssimplicity,andIthoughtsomethingsimilarcouldbevery useful for IV conversion for current-output DACs. I spent some time figuring out how to do itwith JFETs instead. And the solution in the end was obvious.Using JFETs in this circuit has some key advantages, at least for DAC IV conversion. It eliminates allthe biasing resistors and capacitor around the base of the bipolar transistors, and you only need totie both JFET gates to the DAC ground, thus simplifying the circuit even further. All is left now are apairofcomplementaryJFETswithmatchedIdss,a(battery)floatingpowersupply,2decouplingcaps,andoneIV-conversionresistor.Itallseemsobvious,buttheelimination,oratleastdrasticreduction,of the base current through the BJT’s back to the DAC Gnd is a key performance advantage of thisJFET based circuit compared to the srcinal, as will be explained later.IthennotedthesimilarityofthiscircuittoNelsonPass’sZenIV [4] ( Figure1 )intermsofcomponents,thoughtherearealsosomekeydifferencesincircuitry. IntheZenIV,theIVconversionisdonebythetwo 1k resistors at the positive and negative rails, connected to the output via two high-pass filtersin parallel. Thus, the 10uF coupling caps are working in voltage mode, and their non-linearities will1  appearasoutputdistortion.Ontopofthat,anynoiseoneitherpowersupplyrailwillbefeddirectlythrough to the output, i.e. there is no PSRR. This adds to the requirement for the power supply.As I pointed out at the beginning of the Zen IV thread [5] , there are a few more limitations of the Zencircuit. Let’s take a DAC like PCM1704 with 1.2mA output. To get 2Vrms out, one needs to use 4.7k resistorsatbothrailsforIVconversion. Andinordertokeepthecurrentswingsmallcomparedtobias(forlowdistortion),thebiascurrent,inthiscasealsoIdss,oftheJFETswantstobe,say,6mAminimum. Thismeans,however,28Vacrossthe4.7kresistors,requiring+/-37Vrailsorhigher. Theresistorsseecontinuous dissipation of 170mW, increasing thermal noise in the resistors. The equivalent input resistance of the Zen IV equals to the reciprocal of the sum of transconduc-tancesoftheJFETpair,andisabout17ohm. Ifonewishestokeepthislower,tosay<10R,oneneedsto use 2 pairs of complementary JFETs in parallel, each with 6mA+ Idss. This, however, means +/-65V rails, and 0.7W continuous dissipation per resistor.MostofthesedrawbacksarenotapplicabletotheJFET“Leach”circuit,whichInicknamedCenIV(toreflect its complementary nature;  Figure2 ). An 18V supply across the two JFETs is quite sufficient,though you can go further to 27V or 36V if you wish. The benefit of higher voltage is lower JFET ca-pacitances,whichinturnmeanshigherbandwidthandlowerdistortion. TheRivonlycarriesthesig-nalcurrentfromtheDAC,andnoquiescentcurrent. ChangingtheIVconversiongain(fromsay500Rto 2.7k), or increasing the number of JFETs in parallel, does not affect the rest of the circuit – nochanges in rail voltages, or resistor wattage, etc. And the battery only sees a constant current load,2 EUVL Fig1:NelsonPass’ZenIV.  as the DAC current runs in a separate loop and does not go through the batteries at all. The onlydrawbackisthenecessityofafloatingpowersupply. Buttwo9V-blockNiMHrechargableswouldeas-ily provide 10 hours of operation. So this is not a real problem.ImadeadirectcomparisonbetweenthedistortionperformanceoftheZenIVvs.theCenIVinSpice,usingthesameJFETSpicemodelsinbothcircuits. TheZenIVistheoriginalcircuitaspublished,andtheCenIVhasaRivof500RforthesametransimpedancegainastheZen. Theresultsaresomewhatsurprising.Foracurrentinputsignalof+/-1mA,1kHzsine,THDis0.000257%forZen,and0.000066%for Cen. The JFET models being identical, this is pure circuitry difference; a factor of four improve-ment. MyinterpretationisthattheZenIVisapushpulltransimpedanceamplifier,whereastheLeachcircuitisapurecurrentconveyor,asthepowersupplyisnotconnectedtotheDACgroundandthere-forerunsinacurrentlooptotallyseparatefromthesignalcurrentloop. ExceptfortheJFETgatecon-nectionstoDACGnd,theentireinputsignalcurrenthastoreturntoDACGndviaRiv. Thus,whatevernon-linearities there are in the devices, Kirchhoff’s law ensures near-zero distortion (the resistor Rivisalsonot100.000000%distortionfree [6] ). However,becauseofthepresenceofCgdandCgsoftheJFETs,thereisaverytinyleakagecurrentthroughthegatestoDACGndinthepresenceofanACsig-nal, and this tiny current is not flowing though Riv. If this current is not perfectly linearly propor-tional to the input current signal (and it is not), it is precisely this non-linearity that contributes todistortion in the CEN circuit. We shall expect that distortion decreases with JFET capacitances, andincreaseswithfrequency. Whataboutahighertransimpedancethan500R? UsingRivof2.7kintheCencircuit,THDis0.0019%at1kHz. ThefloatingPSUessentiallyswingstogetherwiththeoutputsig-nal relative to DAC Gnd. And the higher the output swing, the higher the leakage current throughthe JFET gates, and thus the higher the distortion. But still, the performance is very respectable byany standard.I could have stopped here. But 2SJ74s are becoming increasingly difficult to get. And the 2J103 /2SK246 pair is of little use here due to its low transconductance. So I wanted to come up with a de-sign with only N-JFETs, and at the same time make use of their lower capacitance to reduce distor-3 Zen  Cen  Sen - Evolution of a Minimalistic IV Conversion Circuit Q12SK170Q22SJ74 V118C1220uFC2220uFC30.1uFC40.1uFR1560kR2560k +Out 2.7kI1I_in DAC 330pF Riv Civ Fig2:TheCenIV (2SK170/2SJ74,matchedIdss)  tionfurther. Thenewcircuit,whichInamedSenIV(SforSingleEnded; Figure3 ),ismuchakintothecurrent-source-loaded JFET follower, published by Curl, Borbely, Feucht, and Pass on various occa-sions.Togetclosetothesameinputimpedanceasthe2SK170/2SJ74pair,atotalof4x2SK170’sareused,asthetwoloweronesmerelyactsascurrentsourceandplaynopartindeterminingtheinputimpedance. Onepenaltyisincreasedcurrentconsumption. Butevenatabout20mAtotalbias,thisis still not unacceptably high. Distortion now is 0.0012% (-98.4dB) with 2.7k Riv, a touch lower thanCen. As already mentioned, only the top two 2SK170’s are contributing to gate leakage related dis-tortion. Sincethe2SK170haslowercapacitanceasthe2SJ74,itisnottoosurprisingthatthedistor-tion in the Sen IV is further reduced. TheseTHDfiguresdonotnecessarilyrepresentreality,astheyareonlyasaccurateastheSpicemod-els themselves, and all passive components, current sources, power supplies are assumed ideal. Itdoes, however, indicate that the development of the circuitry itself is going in the right direction. Bread-boarding&FirstMeasurements Once I was happy with the circuit, it was time to build and test. Idss-matched devices (8.5mA) weredugoutfrommyJFETstockandbreadboardcircuitswerequicklyassembled. Twoimportantissuesbecame obvious at this point – a true floating supply with no coupling to Ground (I used 2x 9V bat-teries which gave no problems at all), and the need of a potential divider to set the floating powersupplysymmetricallyaboutInputGroundduringstartup. Thelatterwaseasilyachievedbyaddinga pair of resistors of identical values in parallel to the output coupling caps. I tried anything from68k to 560k, and they all worked fine, so the value is not critical. Eventually I settled on 560k.Formeasurementpurposes,thetwodecouplingcapacitorsarenotsocritical. Ijustusedapairofor-dinary electrolytic capacitors, such as Panasonic FC 100uF 16V. There is room for experiment in theeventual application, where I plan to start off using German-made Frolyt Bipolar (220uF 25V), fol-lowed by ELNA RBP 220uF 16V bipolar, and Nichicon Muse ES 100uF 16V bipolar, in parallel withWIMAMKP20.1uF250V. Onecouldalsouselarge-valuefilmcaps,suchasWIMAMKS2-XL15uF. Theminimum value of the capacitance should not be less than 4.7uF, assuming a Riv of 2.7k at the out-4 EUVL Q22SK170 V318C1220uFC2220uFC30.1uFC40.1uFR1560kR2560k +Out 2.7kQ42SK170Q32SK170Q12SK170I1I_in DAC 330pF RivCiv Fig.3:TheSenIV (4x2SK170,matchedIdss).
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks