By Julian D. Hirsch
TESTING PHONO CARTRIDGES: A test of a phono cartridge can be undertaken with any of several ends in mind. Ultimately, how ever, what we really want to learn is how it sounds, or, more accurately, how records sound when played by that particular cartridge. An obvious impediment to this aim, aside from differences in individual taste, is the large number of possible "test" records one might choose. It seems likely that some where there exists a record to complement the characteristics of each particular cartridge and make it sound or measure "better" than any other. Unless one is prepared to restrict his listening considerably, or else match each record with its optimum cartridge (which some people very nearly do), this type of evaluation leaves much to be desired. What is the alternative? Let us start with the assumption that a pho no cartridge should translate the modulation of a record groove into an exactly analogous electrical voltage. We would really like its electrical output to duplicate the waveform of the signal that drove the cutter head that made the master disc. Unfortunately, fundamental problems limit any cartridge's ability to achieve that goal.
Two distinctly different aspects of a cartridge's operation combine to create its total performance. First, the stylus must precisely follow (trace) the groove, which was cut with a specially shaped, thin-edge stylus capable of inscribing the very rapid undulations in the groove wall associated with high frequencies.
The playback stylus has a somewhat different shape, and is therefore theoretically unable to trace the path of the cutting stylus accurately.
The inevitable geometric differences between the cutter and the player appear in the cartridge's output signal as distortion. Special stylus shapes have been devised that more closely resemble that of the cutting stylus. Although they usually give better high-frequency performance than a simple conical stylus tip, they only approximate the correct shape.
These considerations, plus others, may complicate the life of the record listener, but they need not interfere with the cartridge measurement process. Keep in mind, though, that test records and normal music records are not the same, so the conclusions a tester might reach about a cartridge's performance with test discs might be quite different from your own impressions of the same cartridge playing "real" records.
After the stylus has traced its path, albeit with distorting deviations, its motion must be converted into an electrical voltage. Just how this is done is not really germane to the tester's work (nor should it be to the consumer's purposes), except insofar as the cartridge loading requirements and output voltage affect the interface with the preamplifier or the test instruments. Most cartridge manufacturers specify optimum resistive and capacitive loads for their products, which (in theory, at least) should allow a test laboratory to du plicate the manufacturer's data.
The cartridge designer has to deal separately with the mechanical and electrical aspects of his product's performance, since in magnetic cartridges (other than the moving-coil type) the electrical circuit is manipulated to compensate for response deficiencies in the mechanical circuit. The tester has no need to make this conceptual distinction; he need only follow the manufacturer's recommendations. And so we come down to the actual measurement process, which is largely a matter of playing records and measuring or other wise analyzing the cartridge's output voltage.
Some engineers have used small mechanical vibrators to excite the stylus, thus freeing themselves from the problem of tracing a groove whose shape is incompatible with the stylus shape. Unfortunately, they also free themselves in the process from any relation ship with the real world.
The first rule of cartridge testing should be to remember (and to recite to oneself before every test) that the data obtained describe only what that cartridge did when it played that record under those load and operating conditions. The data can not be extrapolated to show what might happen with different loadings, different tracking forces, or different arms. And the measurements most definitely cannot be taken as any indication of what the same cartridge, under the same conditions, would do when playing a different test record. By "different" I refer primarily to a product of another manufacturer, but it is also true that no two pressings of the "same" model record will give exactly the same results, and after a number of playings any one pressing will show changes, particularly at the highest frequencies. In any case, if a cartridge specification of frequency response or cross talk does not state which test record was used, it is worthless for purposes of comparison with any other cartridge.
FOR our frequency-response and crosstalk measurements, we use the venerable CBS STR 100 record, which has recently been remastered. Its sweep from 40 to 20,000 Hz is synchronized with the chart drive of our General Radio 1521A graphic-level recorder, and the amplified but unequalized cartridge output is connected to the drive circuits of the recorder pen. The cartridge is loaded as recommended by the manufacturer, and the left-channel output is plotted first, giving a frequency-response curve. The chart is then turned back to its beginning and the same process is repeated with the right-channel modulation. This gives a crosstalk plot for the left channel of the cartridge on the same chart. The output of the right channel is then measured in the same manner.
There are inherent differences between the two channels of the STR 100 record, especially in their crosstalk characteristics. When referring to the crosstalk at a specific frequency, we usually average the two readings. A CD-4 cartridge is measured in a similar manner, using JVC test records that sweep from 1,000 to 50,000 Hz.
A number of records could be used to measure cartridge distortion, and we have used most of them at one time or another.
None has been really satisfactory, for a variety of reasons. Perhaps the most serious problem is the difficulty of correlating any measured distortion with anything one might hear when playing a music record. Unlike most electronic components (but like loudspeakers), phono cartridges distort very differently at different frequencies. This makes it relatively easy to find some test record that will favor a particular cartridge to the detriment of its competitors. No cartridge we have ever tested is "best" on every test record.
For years, we used an intermodulation-distortion (IM) record made by RCA (the Model 12-5-39). It had 400- and 4,000-Hz tones re corded in a 4:1 amplitude ratio at peak velocities from about 7 to 27 centimeters per second (cm/sec). A standard IM analyzer could be used to display the percentage of intermodulation distortion of a cartridge output as it played this record. Disregarding the actual numbers one obtained from such a test, it did show how well a cartridge could cope with extremely high mid-range velocities as a function of tracking force. The RCA record was a 78-rpm disc, but we now use the Shure TTR102, which is a 33 1/2-rpm equivalent.
Shure also makes a test record, the TTR103, that contains several different types of test signal. We use it only for a check on high-frequency tracking. Shaped tone bursts of 10.8 kHz at a 270-Hz repetition rate are re corded at four velocity levels from 15 to 30 cm/sec. Using filters to separate the two frequencies, the amplitude of the 270-Hz component in the cartridge output can be expressed as a percentage of the 10.8-kHz amplitude. A cartridge with linear response at 10.8 kHz will not have any significant amount of the repetition-rate frequency in its output. The measurement results appear in graphic form in each cartridge test report.
SQUARE waves would seem to be useful for testing cartridges. A 1,000-Hz square wave should test the flatness of frequency response from below 100 Hz to above 10,000 Hz, as well as the phase shift of the cartridge over the full audio range. Any resonances show up as "ringing" on the wave form, and the amplitude and duration of the ringing are indicators of the effectiveness of any damping in the cartridge. But there is a problem in that it is extremely difficult to cut a true square wave on a record. The actual shape in the groove is a tri angle, with very sharp "points" that require the cutting stylus (as well as the playback stylus) to change direction in literally no time at all. Obviously, real styli cannot do that. Also, the internal resonances in the cutter system can superimpose themselves on the square wave. The CBS STR 112 record, which we use, has a very detectable 40,000-Hz ringing on its 1,000-Hz square wave. Most quality cartridges whose coil inductance rolls off their response above the audio range yield a very nice-looking square wave from this record, but some moving-coil types and electret cartridges reveal the ringing quite clearly. When we first began to use this record, we incorrectly identified the ringing as arising in (and being a problem of) the cartridges, but we were in formed by the various cartridge manufacturers of the true culprit.
The output voltage of a phono cartridge is one of the few specifications that are reason ably standardized. It is measured at a velocity of 3.54 cm/sec per channel at 1,000 Hz (equivalent to a 5-cm/sec lateral velocity). The CBS STR 100 record has these levels recorded separately for each channel.
"No cartridge we have ever tested is 'best' on every test record." At one time, the vertical-tracking angle of the stylus system in phono cartridges was completely nonstandardized. Some years ago, it was found that the most widely used cutting heads made records that were best played back with cartridges having a 15-degree vertical stylus angle. Gradually, that became an industry standard. More recently, it appeared that a slightly larger angle was more compatible with newer cutting systems, and currently 20 degrees is the accepted standard. We mea sure this with the aid of a record issued by CBS some years ago, the STR 160. It contains fifteen bands of 400-Hz tones recorded at different vertical angles from-6 to +43 degrees. The effect of a vertical-tracking discrepancy (error) between the cutting and play back styli is to increase the second-harmonic distortion in the vertical output of a cartridge.
This is tested by connecting the channels in parallel and out of phase to cancel the lateral output. We simplify the process by connecting one channel of the cartridge output to our spectrum analyzer and observing the second harmonic level. A definite minimum will be found when playing the band whose angle most nearly corresponds to that of the play back stylus.
Some of our test records are used purely for listening tests. This is the most convenient way to judge tracking ability. In the mid range, we use a record with 1,000-Hz tones at a 30-cm/sec velocity. The cartridge wave form, if it is displayed on an oscilloscope, reveals peak clipping or other distortions and al lows one to determine whether the situation can be improved by increasing the tracking force. For very low frequencies we use a record containing a 32-Hz tone recorded with such a large amplitude that the groove shape can be seen from a distance of several feet.
The mere ability of a cartridge to stay in this groove is prima-facie evidence of excellent "static" compliance. Neither of these very useful records, incidentally, is currently "in print." A very useful "listening" test record is one issued by the German Hi Fi Institute. One of its bands contains a 300-Hz tone recorded with increasing levels, expressed in terms of their amplitude, from 20 to 100 microns (a mi cron is one-millionth of a meter). This corresponds to velocities from 4 to 19 cm/sec, "A record issued by the German Hi Fi institute ... has been the nemesis of almost every cartridge we have played it with." which may not sound like much in comparison to some of the information that has been published concerning recorded velocities.
Nevertheless, this record has been the nemesis of almost every cartridge we have played it with. A good cartridge will be able to play the 60- or 70-micron level without audible mis tracking, but only a handful have been able to cope with the 100-micron level. We find this record useful for setting the antiskating compensation of a tone arm, since the audible distortion will be the same in both channels when it is set correctly.
One of the most useful tools for evaluating a cartridge is the "Audio Obstacle Course Era III" (or TTR110) issued by Shure Brothers. This has musical selections recorded at levels from "normal" to far higher than will ever be encountered in real commercial musical recordings. Mistracking can be heard easily, in comparison with the sound of the lower velocity portions, and the material covers the entire audio-frequency range. In our experience, very few cartridges can play the entire test portion of this record without sounds of distress. Since it requires no instruments and is readily available from Shure, this record is one of the most convenient devices for evaluating the quality of one's own record-playing system.
No one of these tests is in any sense definitive; but taken together they can present a pretty clear picture of the overall quality of a cartridge. Although tests tell us little, except in the most general way, about the "sound" of a cartridge, they do indicate how well it can be expected to handle a variety of difficult re corded material.
So far, nothing has been said about the tone arm in which the cartridge is mounted for these tests. It has been suggested to me that a single, high-quality tone arm whose merit is universally recognized should be used for all cartridge testing. This is a great idea, except that no such universally acclaimed tone arm exists. If it did (the SME might have qualified some years ago, and we did use it then), the acceptance of such an arm would be temporary at best. When a new, "improved" arm appeared on the market, all the advantages of the standardized arm would disappear. Since the tone arm bears much the same relation ship to the cartridge that the speaker enclosure does to its drivers, the impracticality of such a proposal becomes even more apparent.
Can you imagine testing all speaker drivers in the same "standard" enclosure, regardless of their specific characteristics and individual requirements? For purely practical reasons, we test cartridges in the tone arms of suitable record players that are on hand for testing at the same time. This minimizes the number of times we have to go through a cumbersome cartridge installation (to us, this is one of the most onerous parts of testing any high-fidelity component). To those critics who feel that we have not used the optimum arm (whatever that might be) for any particular cartridge, we plead guilty, and remind them of the astronomical number of possible combinations of arms and cartridges. We can hardly devote a lifetime to playing mix and match in a search for a sonic Holy Grail! Of course, we also listen to a wide variety of records with every cartridge. However, the present subject is testing, not listening; we'll get around to that some other time.
Also see:
TAPE TALK--Theoretical and practical tape problems solved
Source: Stereo Review (USA magazine)