ORE THAN ONEERTICALTRACKING,~NGLE

'as much undistorted information as pos-sible, and among the VPS parameters.we felt should be optimized was VTA.

. Correct VTA for the VPS, in classical the-ory, involves having the playback anglematch the cutting angle (Fig. 1). Physi-cally, VTA is the angle between the sur-face of the record and the line describedby the contact point of the stylus in thegroove and the pivot point of the can-tilever. This relationship is analogous toand originates in the recording process.The theoretical importance of matching

a M. RISCHand BRUCE R. MAIER'

iscwasher Laboratories has re-cently been investigating high-performance phonograph play-

back, and as a result of our investiga-. ns, we have become aware of severalnusual aspects of the vinyl playback

system (VPS). In one of our studies, weere concerned with the microscopically

small dimensions involved where therecord groove is being dragged past theayback stylus, as well as more visible

aspects such as vertical tracking angle(VTA). In our tests, we wished to retrieve

IIIII~Stylus

., RakeI Angle",

I

VTAs between recording and playbackhas been widely expounded, and popu-lar explanations of causes and effectsare generally available, but they shouldbe examined in some detail to fully un-derstand their consequences. Mismatchof playback VTA to the recorded VTA issaid to cause one minor effect: Frequen-cy modulation of the highs present inprogram material in response to any sig-nificant leveL of low-frequency verticalgroove motion. Referring to Fig. 1 again,notice that the vertical-modulation arc

.Members, Research Staff, DiscwasherLaboratories, Columbia, Missouri

21

VERTICALMODULATION

ARC

iActualPivotPoint

Diamond----+l~ 1,:J_--Stylus VERTICAL CONTACT SURFACE

Bottom of Groove

Fig. 1 - The playback system,showing the relationships ofvertical tracking angle (VTA),stylus rake angle (SRA), andvertical modulation arc; sideview through groove.

AUDIO /MARCH 1981

vFig. 2 - Conical stylus tracingerror with vertical groove modulation,showing contact patch or footprintshift from reference position;side view through groove.

shown is characteristic of any vertical(out-of-phase) modulation in the groove.Because of this arc and the resultant for-ward and backward motion along thegroove, the momentary linear cuttingspeed would be varied proportionally tothe amount of horizontal displacementthis vertical motion causes. Any time thisoccurs, any frequencies higher than thedominant vertical-displacement low fre-quency will be modulated by that low-frequency displacement. Of course, allthis is not a problem if the recorded VTAis replicated in playback, whereby thefrequency modulation (FM) is cancelleddue to the vertical-modulation arc beingthe same. The sonic effects of a VTAerror are similar to flutter, heard aswavering of the high frequencies. Thereis an important difference between typi-cal flutter and frequency modulation dueto VTA error, and this is that the FMcauses wavering in relation to the low-frequency content of the program mate-rial whereas flutter is generally a repeti-tive, continuing variance affecting theentire range of program material. It is thistime-dispersive FM distortion that arisesfrom VTA error and causes the theoreti-cal furor of concern over cartridges withhigher than optimal VTAs.

The literature indicates that if one hada cartridge with an inherent VTA higherthan recorded angle, it could be tiltedback (cartridge body allowing, somewon't) and minimum distortion would beobtained. This isn't quite true, sincestylus shape, in addition to VTA, also in-fluences FM distortion.

Conical (or spherical) styl] have a cer-tain built-in amount of FM distortion witheither a lateral or vertical modulation

22

Fig. 3 - Conicalstylus tracing errorwith lateral groovemodulation, showingcontact point(center of contactpatch) skew relativeto cutting stylus;top view.

(Figs. 2 and 3). As shown in Fig 2, dueto the ultimate rounded tip shape of theconical stylus, a significant vertical mod-ulation will cause the contact patch toshift. From the center or reference posi-tion, a stylus drift or lateral thrust aheadand then behind will cause exactly thesame type of distortion as VTA error.Both are slope related and can augmentor tend to cancel depending on the sign(polarity) of the VTA error. The amount ofFM due to the stylus shape will vary inrelation to the stylus radius and the am-plitude of the vertical modulation.

With lateral modulation (Fig. 3),the FM inherent to conicalstyli occurs on both channels

in opposite directions, so the mono(sum) signal would show little or no FMeffect (depending on the accuracy of lat-eral tracking angle). In stereo playback,however, the two channels illustratedwould have high frequencies recordedalong with the lows. The low-frequencymodulations cause the stylus contactareas to skew or twist, thus causing adegradation of high-frequency phasingand stereo imaging. The net result ofthese and other factors is that correctVTA for conical styli will only minimize,not eliminate, the FM distortions.

Other stylus shapes reduce theseproblems because of their smaller con-tact scanning radii. Typically, today'sgenre of elliptical or modified-Shibatastyli have a scanning radius no smallerthan 0.2 mil, while conical styli rangefrom 0.5 mil to 0.8 mil. This smaller ra-dius allows a more consistently definedcontact with the groove, reducing thesetypes of FM distortion.

Our tests indicated elliptical and Shi-bata-type styli have another parameter ofdramatic importance not present withconical styli. This overlooked factor iscalled stylus rake angle or SRA, which isthe angle that the vertical center line ofthe stylus contact patches make with thegroove modulation ridges. Stylus contactareas for different types of styli areshown as black vertical patches at thebottom of Fig. 4. As seen in this figure,the Shibata-type stylus shows the long-est, narrowest contact, while the conicalstylus predictably has a circular contactpatch or footprint. The elliptical andmodified elliptical stylus shapes fall inbetween these extremes. Due to its long,narrow footprint, the Shibata-type stylusis theoretically very sensitive' to positionalchanges in SRA. Any misalignment ofthe footprint relative to the groove modu-lation ridges will cause its vertical foot-print span to increase, resulting in possi-ble losses of very high frequency modu-lation (scanning loss).

Our detailed models showed that thisincrease in the effective scanning radiusdoes not result in a simple, smooth, ef-fective broadening of the stylus footprint.Due to the nature of the tilted, narrowcontact edge, the manner in which thisedge contacts the groove modulations issomewhat nonlinear and more complexas compared to a simple conical stylusshape of comparable radii. Physicalmodeling showed the tracing errorswhich arise with a misaligned Shibata-type stylus ar~ similar to, but greater indistortion level than those of a conicalstylus, and Similarly result in some low-frequency-dependent FM of the decod-ed high frequencies.

We therefore hypothesized seriousgeometric potential for improper SRA ofShibata-type styli. The groove modula-tions can grab at the scanning edges ofthe stylus, torquing or attempting to twistthe stylus which can send vibrationalshocks up the cantilever.

From the modeling of these factors, itwould seem both VTA and SRA shouldbe corrected for optimal playback. Yetthe physical connection of cantilever andstylus causes both of these parametersto vary simultaneously. There is a fixedrelationship between the inherent VTA ofthe cartridge, when set up as recom-mended, and the SRA which resultsfrom the fixe,d stylus chip vs. cantileverattachment.

When considering the propervertical alignment of our testsystem, we had to decide

whether we should align for correct VTA,proper SRA, or some optimal compro-mise position so as to maximize undis-torted information retrieval. Before wecould make that decision, we decided toassess the effect of varying VTA. Ourtest system consists of a Denon direct-drive DP-80 turntable with a modifiedDA-401 tonearm with the ability of in-play rear-pivot-height adjustment via aprecision micrometer. Several differentmoving-coil cartridges were used. in testswith an HA-1000 pre-preamp runninginto a modified- lab reference preamp.The output of this preamp was observedvia a Tektronix 466A Storage Oscillo-scope and/or with a GenRad 2512Spectrum Analyzer. Test data could beplotted on an Easterline-Angus 575 X-Yplotter for reference plots, etc.

Some initial tests were conductedconsisting of spectrum analysis of verti-cal modulation 1M bands (CBS STR-11 2test record) where the vertical angle wasvaried using a moving-coil cartridge hav-ing a modified elliptical stylus. The re-sults of these were somewhat inconclu- .sive as there seemed to be no clear-cutindication of a minimum level of distor-tion at the various angles of playback.There were some subtle and generallyinconsistent shifts in the distortion spec-tra with changes in VTA, although theirSignificance was not determined untilsome time later.

In order to control the variables in ourtest system, cartridge tests were per-formed using the DIN 45-542 VTA testrecord [1], which has bands with varying

"HYPII!R!LLlPTlCAL""ALLIPTIC""FINE LINE"

"L.A.C,"

"STEREOHEDRON"

COMMON.1 mil

SPHERICAL

COMMON.2x.7mll

BIRADIAL ELLIPTICALSHIBATA,

SHIBATA-TYPE

dB-30

Fig. 4 - Groovecontact geometryof four commonstylus shapes.

Bottom: Side viewthrough groovewith black patchesshowing stylus foot-print or contactarea.Top: End-on view

or profile as stylusrests in groove withno sink-in shown.

Fig. 5 - An 1Mdis-tortion vs. VTA varia-tion comparison,showing null pointsin the high and lowfrequency sections.

23AUDIO/MARCH 1981

__ =Iow frequency bands

___ =high frequency bands

-50

~~~\i=I~V\J\ll\I7

VTA. Two groups of bands are involvedwith VTA determination: One is a hiqh-'frequency 1M tone consisting of 1.85kHz and 3.15 kHz with a high-side 1Mproduct of 5 kHz, whereas the other sec-tion is a low-frequency 1M tone consist-ing of 370 Hz and 630 Hz giving a high1M product of1 kHz. Neither 1M productis harmonically related to the base fre-

\ quencies, and thus no masking confu-sion occurs.

Figure 5 is the data plot from this DINrecord using the cartridge with the modi-fied elliptical stylus. Notice the shallow-ness of the high-frequency bands null(point of minimum distortion) comparedto the low-frequency bands null. There isalso a difference in the angle at whichthe null occurs, which tends to hold truefor any stylus shape with an SRA poten-tial. If VTA were the only effect beingmeasured, the 1M distortion nulls for thetwo bands should be very close in slope,shape, and location. We theorized thatthe observed difference in the nulls (Fig.5) was. due to SRA interaction with theshorter wavelengths involved in the high-frequency section of the tests. The shal-lowness of the high-frequency null ismost likely a result of the different SRA-to-VTA relationships between the cuttingsystem and playback cartridge (about25 degrees difference in this cartridge).We feel it was a similar effect in the initialVTA tests that caused the spectral plotof distortion products to show little over-all change on the CBS test record.

Tests with a cartridge having a coni-cal stylus always gave a much closercorrelation between the nulls for the twodifferent frequency sections and tendedto give a deeper null for the high-fre-quency bands than for other (elliptical,Shibata) types of styli. The reason thatnull points for the two frequency sectionsdo not give exactly the same angle anddepth was hypothesized to be due tosome of the tracing distortion mecha-nisms inherent in the conical stylusshape, as discussed earlier.

One means of isolating VTA parame-ters is to use a cartridge with a conicalstylus, which has no SRAbecause of itscircular contact footprint. Figure 6 is aspectral plot of a vertically modulated 1Mtest band (400 Hz and 4 kHz) made us-ing a moving-coil cartridge with a conicalstylus. The vertical tick marks are 10 dBapart, with the top of the graph startingat -25 dB down from the 400-Hz com-

Fig. 6 - Distortionvariation with VTAchange, showingthe increase of dis-tortion componentswith a 4-degreemisalignment inVTA.

Fig. 7 - SRA vs.VTA (or record/playback anglematch), showingthe increase indistortion with mis-aligned SRA, eventhough the record/playback VTA ismatched.

I

'I', I

I :

ponent; the horizontal ticks are 2 kHzapart on a linear frequency scale fromd.c. to 20 kHz. The dark lines shown arethe original distortion components (plussome noise components as the frequen-cy goes up) at a VTA of 16Y2 degrees,which is the angle cut into the record [2).The dotted lines rising up at some pointsrepresent the increase of those distortioncomponents with an increase in the VTAby 4 degrees to 20Y2 degrees. Noticethe increases in the second-order com-ponents of about 5 dB; consistent withdata reported by others [3] who haveperformed tests with conical styli. Thereare a few other locations where the lev-els come up a bit, but no major trends

'24

___ =Correct SRA

____ =Correct VTA

I ,I

III Iii I i III' Ii j I'I!II!iill

are indicated. The major increases thatresult from the 4-degree VTA error areroughly those predicted by theory andpast experimentation. Since the conicalstylus used has essentially no SRA, thedifferences in distortion spectra are dueentirely to playback angle not matchingthe recorded angle.

This experiment points out very clear-ly that if SRA is not a playback variable,proper matching of record and playbackVTA results in lowest playback distor-tion. It must be kept in mind that the in-crease in the second harmonic of 400Hz is due to slope-related waveform dis-tortion, while the increase of the 4-kHzcomponent sideband is due primarily to

V \J \tJ \J)

increased frequency modulation of 4kHz - a more objectionable form of dis-tortion than harmonic distortion.

The next step was to test for distortiondifferences between the optimization ofVTA for proper SRA alignment or for ver-tical-modulation arc matching. At thisstage of our experimentation, we at-tempted some tests using a Shibata-typestylus with a bent cantilever tube to giveodd combinations of SRA-to-vertica,1modulation arc alignment. We were nev-er able to make a satisfactorily "clean"bend due to the Shibata-type- configura-tion and its need for critical vertical(head-on) alignment, but the data pro-duced were intriguing. .

We were, however, fortunate to havein our stock of cartridges a unit deemeddefective due to a stylus misalignment.This cartridge had a modified-Shibatastylus which was slanted more than atypical unit. When properly aligned forSRA, this cartridge was slightly morethan 4 degrees "low" in proper VTAmatch. Figure 7 illustrates the results ofthe correct SRA versus correct VTA ex-periment, with the same basic data dis-playas Fig. 6. A distinct distortion in-crease is shown when SRA is misalignedand the correct VTA match is also made.Compare Figs. 6 and 7, and it will beseen that for an equal degree ofmisalignment, the SRA parameter ismost significant in causing a rise in dis-tortion, especially higher order distortionproducts. Notice, too, that when themodified Shibata stylus is correctlyaligned for SRA, distortion products areat lower levels than when VTA is correct-<Iy aligned for the conical stylus.' Thesedata are fairly conclusive regardingwhich parameter is of importance for dif-ferent styli. •

Reported listening testsconcemingVTA alignment have said that as little as1/30 of a degree can make an audibledifference in the clarity of the music, witha higher than optimal misalignmentcausing excess brightness. These re-ports typically do not distinguish be-tween VTA and SRA even when the re-port mentions the existence of SRA: Theresults of our tests indicate that theparameter being optimized in these re-ports was almost undoubtedly SRA. Ourown informal listening tests bear this outas well. When SRA is correctly alignedthe sound quality "locks-in" and the re-trieval of minute details is enhanced.

rable 1- Frequency modulation data for different conditions.

Stylus Type Sum of 2nd-Or-der Sideband(Major Distor-tion Level)

Freq. Deviation at 4 kHz

ModulationConicalConicalConicalConicalShibataShibata'Conical

12.0%11.1 %12.8%16.4%

2.1 %1.8%2.2%2.5%1.4%

L 1%, R 0.6%L 1.3%, R 0.8%Mono 0.0% for

both

VerticalVerticalVerticalVerticalVertical'LateralLateral

14016 Y2 019021 Y2°

~14°14016 Y2 0

1CBS STR 112 test record, Band.Z, Side B (400 Hz and 4 kl-lz).

These conclusions were further con-firmed by some tests utilizing a laterallymodulated 500-Hz asymmetrical squarewave cut from Denon test record XG-7003. This recorded signal has a seriesof finely detailed harmonics extendingabove 40 kHz. We postulated that amisalignment of the stylus would alter orlose the harmonics. When a 4-degree tiltto the optimal SRA was introduced, al-teration of the harmonics as low as 5kHz and 7 kHz occurred and losses ofharmonics above 30 kHz were evident!These changes are subtle, but at thesame time consistent and repeatable.

We studied the frequency deviationfor the 4-kHz component of the 1Mtoneused throughout these tests. An experi-mental comparator based on a PLL ICwas used for these tests. While absoluteaccuracy may not hold, the relative rank-ings remain accurate. Table I lists the re-sults of these measurements taken un-der various conditions of VTA. It can beseen from these figures that there is analarming amount of distortion present,although in practice vertical modulationtends to be rare in recording. In fact, thethickness of the recording lacquer, com-mercial considerations, and engineeringexpertise generally keep vertical cuttinglow, and thus phase informationcoherent. .

Our calculations indicate that maxi-mum cantilever vertical-arc travel is typi-cally 1 degree due to these limitations.Another theoretical aspect of VTA matchoften overlooked is cantilever length,which should be matched between cut-ting and playback systems. There is no

standard for these lengths,. and car-tridges we have examined show grossdifferences in cantilever length and donot correlate to cutting systems.

Thus, our investigations clearly showSRA to bea more important controllablevariable than VTA. Our dialogue withcutting engineers indicates that VTA cur-rently varies between 16 and 22 de-grees, depending on the lathe system.SRA, however, is generally 91 to 95 de-grees relative to the record surface in or-der to facilitate lacquer "chip" (cutawaystrand) removal.

Proper hi-fi set-up should thereforeconcentrate on cartridge' adjustment thathas the tip of the stylus pointed "back"toward the tonearm pivot, and the top ofthe. stylus tipped "forward" so that thecontact SRA face is 92 degrees be-·tween the stylus and the record surface.Such alignment will at least approximatecorrect SRA. (One cautionary note: TrueShibata styli do not have their stylus con-tact area or footprint lined up with thebulk of the stylus chip, and this should'be taken into account ~hen adjusting forproper SRA.) The effects are clearly au-dible on a fine audio system. .til

References1. Available from Gotham Audio Corp.. 74 t wasn-ington St., New York, N.Y. 10014.2. White, James V. and Arthur J. Gust, "Three FMMethods for MeasuringJracking Angles of PhonoPickups," Jour. of the Audio Eng. Soc, Vol. 27, No.4, April, 1979, p. 242.3. Halter, Jerome B. and J.G. WOOdward, "VerticalTracking Angle Errors in Stereodisk Systems," Jour.of the Audio Eng. Soc., Vol. 12, No.1, Jan., 1964,.p.8.

25AUDIO /MARCH 1981