EQUIPMENT TEST REPORTS (March 1978)--Hirsch-Houck Laboratory test results on the Spectro Acoustics 210 stereo graphic equalizer. Sony PS-X7 record player, Dynavector 10A phono cartridge, Phase Linear Phase III speaker system, and Soundcraftsmen MA5002 power amplifier.
By Hirsch-Houck Laboratories
Spectro Acoustics 210 Stereo Graphic Equalizer
THE Spectro Acoustics Model 210 is a versatile, high-performance, stereo octave-band graphic equalizer intended for the amateur recordist as well as the serious music listener. Its vertical slider controls (one set for each channel) operate in ten frequency bands, with center frequencies from 30 to 15,360 Hz.
A horizontal slider UNITY GAIN control for each channel permits compensation for any signal loss or gain introduced by the equalization process.
In addition to these sliders, the only controls on the front panel of the Model 210 are five pushbuttons. One is the power switch, with a red pilot LED above it. The next, EQ BYPASS, replaces the equalizer circuits with a direct-wire connection; the system will operate in this mode even with the Model 210 turned off. The EQ LINE button inserts the equalizer into the system. Like most such accessories, the Model 210 is meant to be connected into a system's tape-monitoring path, so tape-recording outputs and playback inputs are duplicated on its rear panel. In the normal mode of operation (EQ LINE), the tape deck receives an unmodified signal. Pressing EQ TAPE includes the equalizer in the signal path going to the tape recorder while the line out puts carry an unmodified signal. The final but ton, TAPE MON, duplicates the functions of the system's tape-monitor switch. It is also possible to connect the Model 210 between a preamplifier and power amplifier, but this is usually less convenient for tape-recording purposes.
With appropriate control settings, the equalizer can, in effect, be placed either ahead of or after a tape deck when recording or playing back a program. The program can be monitored from the tape-playback outputs or at the recorder input. With the TAPE MON switch set to SOURCE, the equalizer operates only on the signal that is passing through the amplifier.
The filter circuits of the Model 210 use a "gyrator" design that electronically simulates an inductor. According to the manufacturer, such circuits have lower noise and hum pick up, more headroom, and lower cost than actual coil inductors. Each of the ten octave-band controls has a nominal range of ±13 dB (±15 dB if two adjacent controls are operated together) around the center (flat) condition.
With the UNITY GAIN controls centered, the gain is nominally 1, but it can be adjusted over a ±15-dB range by these controls. The Spectro Acoustics 210 has a very high output capability, delivering up to 10 volts rms to a 10,000-ohm load without clipping. It can also drive a 600-ohm load with no degradation of performance except a slightly lower maximum-output voltage. This would permit it to drive at least ten power amplifiers simultaneously, should any user wish to do so.
The Spectro Acoustics 210 is finished in black with gold panel and knob markings. It measures 6 x 6 1/2 x 17 inches and weighs 12 pounds. A rack-mounting version with a 19-inch panel is also available. Price (either version): $295. Optional walnut cabinet for the standard model: $40.
Laboratory Measurements. The center frequencies of the individual filters conformed closely to the frequencies marked on the front panel, and the maximum boost or cut provided by each filter was between 11.5 and 13.5 dB, depending on the frequency. When an adjacent pair of filters was used, the maximum boost or cut was about 14 dB at the rated filter frequencies, with a "valley" of approximately 4 dB between them.
The output with a 1,000-Hz test signal clipped at 10.9 volts rms into a high-impedance load and at 6 volts into a 600-ohm load.
The distortion at 1,000 Hz was essentially the residual of our test instruments, reaching a maximum of 0.0034 percent at 10 volts, just before the output clipped. The distortion into 600 ohms, within the maximum output capability at that impedance, was the same as into a high impedance. At 15,000 Hz the distortion was slightly higher, reading 0.004 to 0.007 percent from 0.1 to 7 volts output and 0.14 percent at 10 volts. The IM distortion reading was the 0.002 percent test-instrument residual at all levels up to 10 volts. The UNITY GAIN control's range was from -12.8 to +13.5 dB at 1,000 Hz.
The output noise was measured in a wide-band condition (no weighting). With all the octave sliders centered, the noise level was 225 microvolts (-73 dB referred to 1 volt).
The hum, essentially all 120 Hz, was 81 dB below 1 volt. All these measurements were based on a meter bandwidth of more than 6 MHz; a noise measurement limited to a 20-kHz bandwidth would have provided even better figures.
Comment. The Spectro Acoustics Model 210 has nearly unmeasurable noise and distortion, plus an output capability far exceeding ...
------- The maximum effect (in both boost and cut modes) of each
of the 210's ten filters is shown for a wide band test signal such as pink
noise. The maximum boost or cut averages about 12.5 dB.
...normal signal requirements. As the instruction manual points out, it would be very difficult indeed to create a distorted condition, no matter what associated equipment was used or how carelessly the various equalization controls were set.
The instruction manual is a very well writ ten twenty-eight-page booklet that covers all aspects of the operation of the Model 210. It is evident that Spectro Acoustics feels that for most purposes the adjustment of the equalizer is best done by ear rather than with a pink-noise signal and test instruments. Rather than attempting to equalize room response at a particular point in space, they suggest using the Model 210 as a "creative equalizer" to make the sound suit one's own taste.
There is nothing wrong with this approach, but it can take some practice, since the equalizer has twenty more or less independent adjustments. We have used octave equalizers in the past to reduce the effects of room resonances (aggravated by speaker resonances).
They are invaluable for that purpose, and, by any standard, the Spectro Acoustics Model 210 is one of the best of the genre.
====================
Sony PS-X7 is a fully automatic, turntable
THE Sony PS-X7 is a fully automatic, single-play record player with, among other features, a new direct-drive quartz-locked motor and a carbon-fiber tone arm. In line with a current trend in high-quality turntables, the Sony PS-X7 phase locks the speed of the d.c. servomotor to the frequency of a quartz-crystal oscillator. The turntable speeds of 33 1/3 or 45 rpm are selected by pushbuttons on the motorboard and are not adjustable.
The PS-X7 has a permanently magnetized pulse pattern in a ferric strip around the inside of the turntable rim. A sensing head on the motorboard picks up these pulses as they pass by and generates the feedback control signal that governs the motor speed. The motor has a very high torque so that it can come up to speed (or change speed) very rapidly. It is also totally unaffected by added vertical loads such as those from permanently installed record-cleaning accessories.
Stroboscope markings for the two speeds are cast into the rim of the platter; they are illuminated by a neon lamp next to the platter.
Since there are no speed adjustments to be made, the stroboscope's sole function is to reassure the user that all is well. The PS-X7's record mat consists of two rubber discs sealed together with a viscous fluid inside. This is in tended to minimize vibration transmitted from the turntable platter to the record and to damp any vibrations in the disc itself. The combined weight of the platter and mat is approximately 4 pounds.
The tone arm, conventional in its basic design, is made of carbon fiber. The aluminum cartridge shell, of the widely used bayonet-mount type, is cut away to minimize its mass.
The counterweight carries the tracking-force scale, which reads from 0 to 3 grams in 0.1-gram intervals. An antiskating dial next to the arm's base is calibrated over the same range.
The cueing lever is part of the arm base and operates through a curved bar that rises under the arm tube.
A four-position knob control on the motor-board has automatic arm-indexing settings for 17-, 25-, and 30-cm (7-, 10-, and 12-inch) records, plus a MANUAL setting. The other operating controls of the PS-X7 are on the front of the base and are thus accessible with the dust cover closed. A pushbutton turns the power on and off, and there are electronic touch controls for automatic START/STOP and REPEAT functions. The base itself is molded of Sony's SBMC plastic compound, which looks and feels like cast metal but is claimed to be acoustically inert.
Touching the START/STOP contacts turns the motor on and causes the tone arm to index to the selected diameter and then lower to the record surface. (In the MANUAL mode, the motor will start but the arm must be cued in manually.) Touching these contacts again causes the arm to lift and return to its rest, after which the motor shuts off. At the end of a record side, a photo sensor under the motor-board automatically trips the shut-off mechanism without exerting any force on the tone-arm system. To repeat a record, one touches the second pair of contacts (REPEAT); the record will then play repetitively until either REPEAT Or START/STOP is touched again. Each pair of contacts has a LED between them that glows when the function is activated.
The Sony PS-X7 is mounted on very resilient feet, viscous-damped like the record mat, to isolate it from external vibration and shock, and it is furnished with a hinged plastic cover. The signal-cable capacitance is compatible with the requirements of CD-4 cartridges. The record player is 17 1/2 inches 6 inches high, and 14 3/4 inches deep; it weighs 22 1/2 pounds. Price: $330.
Laboratory Measurements. We installed a Shure M95ED cartridge in the head shell of the PS-X7 for our tests. The record player comes with a protractor for accurate setting of the stylus overhang, but we found that the recommended spacing of 49 mm from the stylus to the reference mounting surface of the head shell was sufficiently accurate to give a low tracking error.
The arm's maximum tracking error was an excellent 0.5 degree per inch of radius, and the typical error was a fraction of that amount. When the arm was balanced according to the instructions, the tracking force was about 10 percent higher than indicated, and the vertical balance condition was somewhat ambiguous. We therefore suggest using an external stylus-force gauge for operation at 1 gram or less.
The automatic operation of the PS-X7 could hardly be simpler. A touch of the finger on the START/STOP contacts is all that is required. The arm is always raised when the unit is shut off, regardless of the position of the cueing lever. This makes the stylus relatively immune to damage from accidental displacement of the tonearm. However, it does complicate checking the tracking force or the tracking error.
The quartz-locked turntable, as might be expected, operated at exactly correct speeds, which did not vary with line-voltage changes or even rather large mechanical drag forces applied to the platter. The torque of this turntable is so high that one must literally grab the platter and stall it in order to make any perceptible uncontrolled alteration in its speed.
On the other hand, shifting between 33 1/3 and 45 rpm took less than 0.5 second. The un weighted rms wow was 0.05 percent, and the flutter was 0.03 percent (this figure cannot be compared to the manufacturer's rating of 0.025 percent, which is a weighted measurement). The flutter spectrum had a broad maximum in the 50- to 70-Hz range, which would be attenuated significantly by any of the currently used weighting curves.
The rumble of the PS-X7 was lower than average, even for a direct-drive turntable.
Unweighted, it was-35 dB, and with ARLL weighting it was-62 dB. The rumble was principally at frequencies below 30 Hz. Out of curiosity, we substituted a rubber mat of conventional construction for Sony's fluid-filled mat. It made no difference in the rumble-or anything else-that we could measure.
The arm mass was slightly lower, at 17 grams, than we have measured on many other automatic record players. It resonated with the compliance of the M95ED stylus at 8 Hz with an amplitude of 6 dB. The cable capacitance was 90 picofarads (pF) from either channel to ground and a low 3.3 pF between channels. The anti-skating was optimum when set to match the tracking force. Cueing had to be done slowly to avoid jarring the pickup. On its descent, the arm drifted slightly outward, repeating several seconds of the record. The automatic operating cycle was quite fast, with 8 seconds elapsing between touching START and the set-down of the stylus in the lead-in groove. The shut-down cycle required only 4 seconds.
The soft mounting feet give the PS-X7 exceptionally good isolation from base-conduct ed vibration. Its susceptibility to such vibrations from 20 Hz on up was typically 10 to 20 dB lower than that of most comparable record players we have tested. In fact, only three or four players we have handled over the years have had measurably better isolation than the PS-X7.
Comment. The externally accessible operating controls of the PS-X7 are undeniably convenient, but they go only part of the way toward permitting complete operation of the record player with the cover down. If cueing could also be done from the front, the ma chine's utility would be greatly enhanced.
Moreover, it seems to us that in the details of its design the tone arm of the PS-X7 does not match the refinement exhibited by the turntable itself. Aside from the already mentioned tracking-force ambiguity and cueing drift, we found the finger lift poorly designed.
However, these few criticisms are not in tended to denigrate the overall quality of the PS-X7, but only to put matters into perspective. For its basic function-playing records-this machine would be hard to beat.
The acoustic isolation of the base of the PS-X7 is far superior to that of the typical direct-drive record player, and we would rate that feature as more important than all the minutiae of tracking error, rumble, and the like, for these have little or no audible effect once a certain level of performance has been surpassed. Acoustic feedback, in contrast, is probably the most serious-and, usually, least appreciated-of the problems afflicting record players today. In many installations, the 20 dB or so of added isolation afforded by the PS-X7's mounting feet could well be crucial.
The special magnetic pickup head of the PS-X7 is seen positioned on the base near the edge of the platter. The magnetic strip on the platter edge is shown at right. It has "prerecorded" on it a series of precisely spaced pulses (like a length of recording tape). The pickup senses the pulse-repetition rate and uses it to monitor and control (via a phase-locked loop) the platter's speed.
=====================
DYNAVECTOR 10A phono cartridge
DYNAVECTOR moving-coil phono cartridges are manufactured in Japan by Onlife Research, Inc., and distributed in this country by Audioanalyst, Inc. (Elsewhere they are better known as "Ultimo" cartridges.) Like most moving-coil cartridges, they have styli that cannot be replaced by the user. Unlike most, they have a relatively high signal output (nominally 2 millivolts), which permits connecting them directly to most preamplifier phono inputs without interposing a "pre-preamplifier" or step-up transformer.
There are several cartridges in the Dynavector line, all similar except for their stylus assemblies. The two most expensive models (20A and 20B) have Shibata styli mounted, respectively, on aluminum and beryllium cantilevers; they operate at a vertical tracking force of 1.5 grams. Their high-frequency response extends to 50,000 Hz, making them appropriate for CD-4 quadraphonic playback as well as conventional stereo.
The Dynavector 10A, which is the subject of this report, is the least expensive cartridge in the Dynavector line. It employs a 0.6-mil spherical stylus on an aluminum cantilever.
The stylus assembly is also "stiffer" than the assemblies of the 20A and 20B and requires a tracking force of 2.5 grams.
The other specifications of the 10A cartridge are the same as those of the more ex pensive 20 series, including an 85-ohm d.c. coil resistance, 2 millivolts output, and more than 20 dB of channel separation at 1,000 Hz.
As with other moving-coil cartridges, the 10A's frequency response does not depend on the 10A is not a CD-4 cartridge, but it is virtually equivalent to the 20 series in stereo reproduction.
The Dynavector 10A is housed in a translucent red-plastic body with standard 1/2-inch mounting centers. The cartridge mass of 9.5 grams can be balanced by almost any tone arm. The connection pins on the cartridge have a smaller diameter than those used on other cartridges, but a set of connecting wires fitted with smaller clips is supplied with the cartridge to simplify its installation in tonearm head shells. Price: approximately $165.
Laboratory Measurements. The Dynavector 10A was tested while mounted in the tone arm of a Sony PS-X5 record player. We also installed it in several other record players for listening tests. Since the cartridge instructions are quite specific in stating that the tracking force should be 2.5 ±0.1 grams, that is the force we used. The cartridge output was al most exactly 2 millivolts, as rated, for a re corded velocity of 3.45 centimeters per second (cm/sec), and the channel levels were balanced within 0.65 dB. The vertical stylus angle was 22 degrees, close to the industry standard of 20 degrees.
The 10A tracked our high-velocity test records, including the 70-micron level of the Ger man Hi Fi Institute record, without difficulty.
The intermodulation distortion as measured with the Shure TTR-102 record was between 2 and 3 percent at velocities of up to 23 cm/sec, and 7 percent at 27 cm/sec. This is good performance, though not exceptional for a high-quality cartridge. However, when we played the 10.8-kHz tone bursts of the Shure TTR-I03 record, we found that the repetition-rate distortion was at least as low as we have ever measured and considerably less than the typical readings of top-ranking cartridges. It varied only slightly with level, from 0.6 percent at 15 cm/sec to under 1 percent at 30 cm/sec.
We measured the 10A's frequency response with the CBS STR 100 test record.
Cartridge loads from 100 to 100,000 ohms had no effect on the response curve, although the 100-ohm load dropped the output level some what. The frequency response was smooth and flat throughout the audio range to 5,000 Hz, rising gradually to +4.5 dB at 20,000 Hz.
(A response curve provided with our test sample-one is included with every Dynavector cartridge sold-showed a rise of only 2 dB at 20,000 Hz; however, it was made using a B&K QR-2009 test record and thus cannot be compared with our results from the CBS record.) Channel separation was from 20 to 30 dB over much of the audio range, measuring from 15 to 20 dB at 10,000 Hz and from 10 to 13 dB at 20,000 Hz.
With the Sony arm, the low-frequency resonance took the form of a double peak at 12 and 18 Hz, and there was less than I dB rise in response at any frequency above 30 Hz. The square-wave response (measured with the CBS STR 112 record) was nearly perfect, with only a single overshoot whose amplitude was about 70 percent of the square-wave amplitude. There was the usual 40,000-Hz ringing across the square wave, a property of the record. The shape of this 1,000-Hz square wave tells us not only that the cartridge response is flat from below 100 Hz to above 10,000 Hz, which we knew from other measurements, but also that it introduces negligible phase shift over that range.
In listening tests with the Shure "Audio Obstacle Course-Era III" test record, the Dynavector 10A was most impressive. Not only did it play everything on the record without audible mistracking, but it reproduced the maximum levels of the sibilants and the musical bells without the slightest sense of strain.
Except for being louder, they were indistinguishable in playback from the lowest-level passages of these sounds on the recording.
This is unusual, since even cartridges that do not obviously mistrack generally sound a trifle strained at the maximum levels.
We were concerned with the need to operate the cartridge at 2.5 grams. Exerted through a 0.6-mil stylus, this force could conceivably cause undue record wear. To settle our doubts on this matter, we played a three-minute portion of a very high quality record some 120 times in succession. We expected that if there were any significant amount of wear, we would hear some degradation in that portion of the record compared with the preceding and following sections. There was not. We could not tell, either by listening or visual examination, where the groove had been played only once and where it had been played 120 times. On the assumption that no single record is likely to be played more than 100 times during an owner's lifetime (and certainly not in succession without some cooling-off period), we concluded that, at least with this particular cartridge, the 2.5-gram force does not represent a threat to one's record collection.
Comment. We are well aware of the reputation some moving-coil cartridges have for listening qualities that do not necessarily correlate with their measured performance.
In the case of the Dynavector 10A, the correlation was very close and we were not compelled to ascribe its sound quality to any un known or mystical factors.
And that sound was, in a word, superb.
This came as no surprise, since the 10A (at its 2.5-gram force, to be sure) had better tracking ability throughout the audio range, and especially at the highest frequencies, than just about any other cartridge we have used. Its frequency response was in general very flat, rising only at the extreme top end where a little added "bite" is often beneficial. The cartridge is exceptionally free of phase shift in the audio range, if that matters (we are not sure it does). The response, both in and out of the audio band, is completely independent of external load conditions, and the cartridge it self will not affect the equalization of any phono preamplifier. The cartridge's low impedance will in most cases significantly re duce preamplifier input-circuit noise. Finally, the low stylus compliance places the arm resonance well above the region where warp effects could degrade the sound quality.
------ In the graph at left, the upper curve represents the smoothed, aver aged frequency response of the cartridges right and left channels; the distance (calibrated in decibels) between it and the lower curve represents the separation between the two channels. The inset oscilloscope photo shows the cartridge's response to a recorded 1,000-Hz square wave (see text), which indicates resonances and overall frequency response. At right is the cartridge's response to the intermodulation-distortion (IM) and 10.8-kHz tone-burst test bands of the TTR-102 and TTR-103 test records. These high velocities provide a severe test of a phono cartridge's performance. The intermodulation-distortion (IM) readings for any given cartridge can vary widely, depending on the particular IM test record used. The actual distortion figure measured is not as important as the maximum velocity the cartridge is able to track before a sudden and radical increase in distortion takes place. There are very few commercial phonograph discs that embody musical audio signals with recorded velocities much higher than about 15 cm/sec.---
Impressive as these attributes are, we would be somewhat less enthusiastic if they were attainable only by using a $200 cartridge with a pre-preamplifier of comparable cost, which is the case with most other moving-coil cartridges. But (and it is a vital distinction) the 10A costs only slightly more than the best conventional magnetic cartridges from several major manufacturers and its output is compatible with any good amplifier without requiring a further investment in step-up or amplifying accessories. Even if much of the time its sound is indistinguishable from that of a similarly priced conventional cartridge in a listening test, there will be some occasions when it will outperform the latter.
Its only drawbacks, so far as we could determine, are the non-replaceable stylus (which is no problem if one exercises care in the installation and use of the cartridge and keeps it away from small children) and the fact that its mounting holes are about 1/2 inch closer to the stylus than those of virtually every other cartridge we have tested in the past twenty-odd years. This means that in some tone arms (Dual, for example) there will not be enough overhang-adjustment range for minimum tracking error.
If you can buy, install, and use the Dynavector WA and manage to overcome any pre conceptions about the relationship between tracking force and record wear (which was one of the most difficult problems for us), the sound will more than compensate for any minor inconveniences. We understand that a less expensive model, the 10X, will be avail able shortly.
===================
Phase Linear Phase III Speaker System
THE Phase III loudspeaker system is the latest product to emerge from the innovative mind of Phase Linear's founder, Bob Carver. Like his noise-reducing "Autocorrelator" and the largest super-power amplifiers for home use, the Phase III departs visibly and audibly from common practice.
Like that of most speaker designers, Carver's goal was to generate a smooth, wide-range, essentially undistorted response over the full audible spectrum. Beyond that goal, he also hoped to re-create at the listener's ears a sound field that would resemble the conditions existing in a concert-hall environment and therefore be perceived as sounding more "realistic" than conventional loud speaker sound.
Carver's investigations into human perception of sound led him to the conclusion that four basic objective/subjective factors must be considered: Echo, or reverberation, gives the listener a sense of the spatial properties (size and acoustic characteristics) of a concert hall or listening room. The phase relations of perceived sounds reaching one's ears at different times help locate sound sources in space. Short-term reflections (too close in time to the original sound and to each other to sound like echoes) add a sense of depth. And finally, there is the difference in the energy spectrum reaching one's two ears arising from interference effects between different frequencies as they travel slightly different paths. (Carver refers to this as the "bounded-ripple" or "comb-filter" effect.) It is Carver's belief that a correct mixture of these four properties will enable a listener to experience more believable (or realistic) music reproduction in his home, and his reasoning is present ed in detail in the pamphlet describing the theory behind the novel design. We, of course, were more concerned with how this theory was translated into hardware, and were willing to let the Phase III (originally Andromeda III) speak for itself, as it were.
The Phase Linear Andromeda III was de signed to be as omnidirectional as possible. It consists of two large upright panels and a separate bass module ("sub-woofer") together with an active equalizer that is called, rather cryptically, a "Motion Control Module." Each panel contains two 8-inch upper-bass drivers facing forward but open to the rear (so that their radiation is in a dipolar or "figure-8" pattern). These drivers operate from approximately 100 to 600 Hz. Higher up on the panel are two small cone mid-range drivers, nominally 4 inches in diameter. One faces forward and one to the rear, and they operate from 600 to 3,000 Hz.
Surrounding the mid-range drivers are four small cone tweeters (approximately 1 inch in diameter) mounted in cut-outs in the panel and facing upward. A plastic conical deflector above each tweeter gives it a nearly omnidirectional response at frequencies above 3,000 Hz. In the center of the panel is a 1-inch dome tweeter angled slightly upward. A crossover network on the back of the panel distributes the signal to the various drivers. The speaker panels, which have removable cloth grilles on both sides, have metal supporting feet that angle the entire panel slightly backward.
Although the 8-inch woofers (which have compliant, long-throw suspensions) would be capable of excellent bass response, the size of the baffle on which they are mounted limits their effective range to frequencies above about 100 Hz. The lower frequencies for each channel are radiated from separate 12-inch drivers housed in a single internally divided enclosure with two ducted ports. The sub-woofer has been designed to provide a strong, low-distortion bass output down to 24 Hz.
The signal leads from the amplifier go directly to the bass module, where the first crossover takes place, and the frequencies above 100 Hz are channeled from there to the two panels.
The electronic "Motion Control Module" is actually an equalizer, a unity-gain device meant to be connected between the preamplifier and the power amplifier. Three slide switches provide a slight adjustment of the levels in the low-, middle-, and high-frequency ranges. Each switch boosts or cuts the level in its range by 2 to 4 dB. A "spatial-imaging" knob gives a continuously variable cross-blending of out-of-phase information between the two channels to provide in creased ambiance.
The upright panels of the Phase Linear Andromeda III are 24 inches wide, 63 inches high, and 5 inches deep; they weigh 70 pounds each. The bass module is 22 inches square and 18 1/2 inches high; it weighs 90 pounds. The system's total of twenty drivers gives it a very large power-handling ability, and it is rated
for operation with amplifiers delivering from 50 to 350 watts per channel. The nominal impedance is 6 ohms. Separate fuses on each panel and the bass module protect the system against damage. Price: about $1,300.
Laboratory Measurements. The large number of dispersed drivers used in the Andromeda III system would make anechoic measurements essentially meaningless be cause of the inevitable interference between the drivers. However, the system should be well suited to the type of live-room measurements we make, in which the total output of the system is measured after any absorption by the normal room environment.
With the system's equalizer switches centered and its spatial-imaging control turned off, we placed our microphone on-axis about 10 feet in front of one speaker panel. It was about 12 feet from the other panel and approximately 30 degrees off its axis. The speakers were placed about 5 feet from the back wall and 3 to 4 feet from the side walls of a 15 x 20-foot room. The bass module was placed near the back wall about midway be tween the speakers, although its output was not included in the live-room measurements.
The smoothed response curve showed moderate dips at 550 and 3,000 Hz, possibly due to crossover effects at these frequencies.
The high-frequency output was strong, measuring somewhat higher than the mid-range level. It was quite uniform between 5,000 and 15,000 Hz (the upper limit of our microphone calibration). The omnidirectional radiation of the speaker panels is illustrated by the fact that the response curves from the left and right panels were virtually identical over the full measurement range.
The woofer response was measured with close microphone spacing to simulate anechoic conditions. The response was measured separately at the driver cone and the port, with the two combined to form a total bass-response curve. Since the bass crossover net work is in the bass module, its effect is included in the measurement. The woofer response peaked at 55 Hz, dropping off above and be low that frequency. Splicing this curve to the high-frequency curve, we obtained a combined frequency response within ±3.5 dB from 30 to 5,000 Hz with an elevated but flat top end about 5 or 6 dB above the mid-range level. With the equalizer set for decreased high-frequency output, the speaker response was within about ±5 dB from 28 to 15,000 Hz.
Because of the unusual nature of the Andromeda III and the considerable contribution of speaker placement (and, to some extent, of the equalizer settings), these numbers should not be given undue weight. What they do show, quite clearly, is that the Andromeda III is a very wide range, nearly omnidirectional system.
The equalizer actually has a very modest effect on the total sound quality of the sys tem. Unlike some speaker-system equalizers, its electrical response with centered controls is nearly flat, within ±2 dB from 20 to 20,000 Hz. The switches have a typical range of about ±3 dB, so that even in the most extreme case there should not be excessive demands on the amplifier's power output be cause of equalizer settings.
------- Tone-burst response of the Phase Ill at (top to bottom) 100, 1,000, and 6,000 Hz in a live environment. The lower trace is the speaker output.
However, the Andromeda III in and of it self is extremely inefficient, requiring more amplifier power than almost any other system we know of. This is not too surprising; amplifier power is the very basis for the existence of Phase Linear, and Carver was much more concerned with achieving certain quality goals with this speaker than with conserving power. We found that 1 watt of random noise in the octave centered at 1,000 Hz produced a sound-pressure level of 80 dB 1 meter from the center of the grille. Driving the sys tem with a Phase Linear 400 amplifier, we frequently noticed the amplifier's meters reaching the 0-dB mark, and when the spatial-imaging control was turned up to blend the out-of-phase signals (which increased the equalizer gain some 5 to 7 dB in the low- and mid-frequency ranges) it was very easy to drive the meters to their limits in spite of the fact that we do not normally listen at ear-splitting levels. We were, perhaps, seduced into using higher levels because the Andromeda III did not sound strained under any conditions we could create.
The Andromeda III presents a very large apparent sound source which never seems to sound as loud as its measured output. We are sure that it could handle the full output of a Phase Linear 700 amplifier, for example, without sounding unpleasant (and it would be very loud under those conditions). For ordinary listening, we found the 200 watts plus per channel of the Phase Linear 400 to be adequate, but we have reservations about the suitability of, say, a 50-watt amplifier for more than background-listening levels.
The bass distortion of the Andromeda III at a drive level of 2.8 volts ( I watt into 8 ohms) was unusually low. It measured between 0.5 and I percent from 100 to 50 Hz, and it never exceeded 2 percent all the way down to 20 Hz. The measurement was made at the cone for frequencies above 40 Hz and at the port for lower frequencies (where its radiation was dominant). A 10-dB increase in drive level produced considerably higher distortion from 1.5 to 3 percent down to 50 Hz and up to 9 percent at 30 Hz.
The overall system impedance was quite low: about 3 ohms at 20 Hz and from 5,000 to 20,000 Hz. It was between 4 and 6 ohms at most frequencies from 30 to 1,000 Hz except at 450 Hz, where it reached its maximum of 12 ohms, and between 2,000 and 3,000 Hz, where it was about 8 ohms. Since no one is likely to parallel two of these systems on a single amplifier, this low impedance will probably not cause problems and will make it easier for any amplifier to deliver its maximum output to the system.
In view of the large number of drivers in the Andromeda III, which could be expected to create interference effects, the tone-burst response of the speaker was surprisingly good at all frequencies. There were no obvious signs of ringing or other anomalies.
Comment. Although we have not emphasized the physical size of this system, it is what first strikes anyone seeing it for the first time. The two panels will dominate any ordi nary-size room.
We listened extensively to the Andromeda III both before and after making our measure ments. Our first impressions remained un changed and were not explained to our entire satisfaction by the test results. The Phase Linear Andromeda III had one of the smooth est, most extended, and best dispersed "top ends" we have ever heard. Most of the time we listened with the equalizer high-frequency switch centered, but sometimes (especially with program material having excellent highs) we preferred a reduced setting.
With this superb high end goes a completely unified sound. One never senses that each panel has nine different drivers; rather, the sound emerges from the entire panel surface somewhat as it does in a full-range electro static speaker (but the Andromeda III sounds very different from such a speaker overall).
However, like an electrostatic (or dipolar or omnidirectional radiator), the panels of the Andromeda III can be approached with no sense of increasing sound pressure. In fact, we spent much of our time within five feet of one of the panels without ever feeling over whelmed or "too close" to the sound source.
The polar response of the panels is actually dipolar over most of the audible range and omnidirectional at other frequencies. To our ears, it came close to sounding completely omnidirectional; we could walk behind the speaker panels with virtually no change in sound quality.
The bass module could never be heard as a distinct or separate contributor to the total sound. It probably could have been placed anywhere in the room with equal success. The low-bass output was most impressive, especially on pipe organ and bass-drum sounds, but with most program material one might never suspect that the sub-woofer was in use.
In fact, at one point we blew the fuse of the bass module and didn't realize it for some time after, since most program material has little energy below 100 Hz. We greatly appreciated that the Andromeda III never added artificial heaviness to male voices, outperforming many otherwise bass-shy speakers in this respect.
We were concerned that our simulated live-vs.-recorded listening test might not be applicable to this speaker. In order for a speaker to accurately imitate the sound of our "live" program, it must have a polar response similar to that of our reference source-and nothing could be more dissimilar in configuration than the Andromeda III and our program source. Not surprisingly, the Andromeda III never sounded so much like the live sound that we were unable to tell them apart. How ever, it did have the same overall frequency balance and we could not hear any change in highs or middles when we switched from the original program to the Andromeda III trying to reproduce it. Only the spatial properties changed, not the sound quality.
While we cannot comment on the validity of the reasoning that led to the creation of the Andromeda III, there is no question that the
-------- One of the Phase Ill's large and complex speaker panels with grille removed and the subwoofer turned on its side to expose drivers.
design provides a sense of spaciousness missing from the output of most conventional speaker systems. One is never conscious of the speakers as the immediate source of the sound, as is usually the case with convention al systems. Yet, though the apparent sound sources within the recorded program were never "pin-pointed" by the Andromeda III, they were also never vague or directionless.
After living with these speakers for some time, we gave up trying to analyze their qualities and just enjoyed the experience of listening to them.
To summarize, Bob Carver's venture into speaker-system design has resulted in a product as innovative as his electronic creations have been, despite their total dissimilarity.
The Andromeda III requires a very large amplifier, of course, but Phase Linear has done more than their share to make that possible. Our capsule impression of the Andromeda III is that it is one of the most visually obvious yet sonically subtle-and impressive-speakers we have experienced. You will have to hear it for yourself, since it sounds like no other we can think of.
===========================
Soundcraftsmen MA5002 Power Amplifier
THE proliferation of new power-amplifier i "classes" continues unabated, with the Soundcraftsmen MA5002 (tentatively labeled "Class H") power amplifier being the latest variant to reach the consumer market. The class of an amplifier circuit usually refers to its internal operating conditions, such as the quiescent (no-signal) current drawn by its tubes or transistors and the fraction of the signal cycle during which each part of the amplifier is active. Recent developments have extended this usage somewhat to include any means by which an amplifier is induced to de liver more power, or to deliver it more efficiently, than conventional design allows.
In the case of the MA5002, the amplifier circuits (except possibly the protective circuits) are conventional and operate in class AB at all times. The difference lies in the use of a power supply whose voltage increases with the signal level, thus giving the amplifier more "headroom" as needed to accommodate high signal amplitudes. Given low- or no-signal conditions, the output transistors operate with a reasonably low voltage and correspondingly low power dissipation. The result is a very powerful amplifier (rated to deliver 250 watts per channel to 8-ohm loads from 20 to 20,000 Hz with less than 0.1 percent total harmonic distortion) that is claimed to draw appreciably less power from the 120-volt a.c. power line when idling or delivering normal listening levels than other amplifiers of com parable ratings. At full power output, how ever, the MA5002 draws an impressive 1,000 watts from the power line.
The Soundcraftsmen amplifier has two power supplies (actually, each is a dual positive/negative voltage supply), with one delivering up to 50 percent more voltage than the other. At power outputs up to 50 percent of the rating (125 watts), only the lower-voltage power supply is connected to the output stages and the unit operates like a very conservatively rated I25-watt amplifier. When the instantaneous signal amplitude is sufficient to drive the amplifier beyond 125 watts, a "Vari-Portional" circuit goes into action and begins to turn on the high-voltage supply, which takes over smoothly from the lower-voltage one. It follows the signal waveform so rapidly that the effective operating voltage on the output transistors increases faster than is required to handle the increased signal level.
As the manufacturer points out, the high-volt age power supply is in effect a powerful, wide-band audio amplifier whose slew rate is greater than that of the audio amplifier proper (which is itself rated at a very high 50 volts per microsecond).
No current limiting is used to protect the output transistors. Instead, the MA5002 has an "electronic crowbar" that shuts down the amplifier totally in the event of an output short circuit or excessive load current. After about 2 seconds, the unit turns on again automatically (if the fault remains, it will not come on fully and will emit a click every 2 seconds).
The front panel of the MA5002 has two large meters calibrated from-20 dB to +3 dB; their 0-dB points are supposed to correspond to the rated output of 250 watts into 8 ohms. Between the meters are pushbutton switches that increase their sensitivity by factors of ten or one hundred so that they can indicate output levels of a fraction of a watt.
Each channel has a front-panel LEVEL control plus a green VARI-PORTIONAL light and a red CLIPPING light. The green light begins to glow when the higher-voltage power supply goes into action, and the red light is a fast-responding indicator of output clipping. An OVER LOAD light comes on when the electronic crowbar has been tripped, and there is a red POWER pilot light. Two SPEAKERS switches and a POWER switch complete the controls.
Most of the rear apron of the MA5002 is de voted to two large heat sinks; between them are two pairs of speaker-output binding-post terminals and the two input phono jacks.
There is a power-line fuse and a heavy-duty line cord with a three-prong grounding plug.
The Soundcraftsmen MA5002 is furnished with finished-walnut side plates, removable for rack mounting, and sturdy black handles on the front panel. The amplifier is 19 inches wide, 15 inches deep, and 7 inches wide; it weighs 55 pounds. Price: $699.
Laboratory Measurements. The one-hour pre-conditioning at one-third rated power made the amplifier's heat sinks too hot to touch for more than a couple of seconds.
However, throughout the rigorous testing that followed there was no sign of excessive temperature on any other external surface of the amplifier.
With a 1,000-Hz test signal, both channels driving 8-ohm loads, clipping occurred at 312.5 watts per channel. The 4-ohm output before clipping was a very impressive 458 watts, and the 16-ohm clipping level was 207 watts. At the rated 250 watts output, the harmonic distortion was well under the 0.1 percent rating, measuring about 0.03 percent at 20 Hz, 0.005 percent at 100 Hz, and increasing smoothly to 0.07 percent at 20,000 Hz. At half and one-tenth power the characteristic was similar, with lower readings.
At 1,000 Hz, the harmonic distortion was between 0.007 and 0.025 percent from 0.1 watt to 310 watts output. The intermodulation (IM) distortion was nearly constant at around 0.05 percent (it ranged from 0.034 to 0.065 percent) over the full power range from less than 10 milliwatts to 310 watts.
The amplifier required an input signal of 0.25 volt for a reference 10-watt output, and the unweighted noise level was 85.4 dB below 10 watts (or 113 dB below rated power). The rise time of the output waveform was 3 microseconds, and the slew rate was 48 volts per microsecond on one channel and 40 volts per microsecond on the other. The frequency response was down 0.2 dB at 5 and 25,000 Hz, and down to-3 dB at 100,000 Hz. We also made a difference-tone distortion measurement using equal-amplitude signals at 18 and 19 kHz and driving the amplifier to a level only 0.2 dB below the clipping point. The 1,000-Hz difference-frequency IM product was 79 dB below either input tone (85 dB be low peak-envelope power), and third-order distortion products at 17 and 20 kHz were at -62 dB.
The VARI-PORTIONAL lights began to glow at 125 watts, and the CLIPPING light flashed at 331 watts. The literature on the MA5002 includes scope photos illustrating the response of the Vari-Portional circuit to high signal levels. We repeated this test for ourselves, and the resulting scope photos show clearly how the power-supply voltage (upper line) begins to increase as the peak of the signal waveform reaches 125 watts, finally increasing enough to accommodate some 380 watts output at the clipping point. (This reading is higher than our original clipping measurement because we drove only one channel and allowed the line voltage to rise to 125 volts instead of the standard 120 volts.) The accompanying photo shows only the positive power supply, but an identical effect takes place with the negative supply voltage.
We short-circuited the amplifier outputs repeatedly, at various power levels, and the crowbar circuit worked perfectly every time to prevent any damage to the amplifier. The output-meter calibration was such that in its Scope photo of the amplified sine wave at the MA5002's output and the power-supply voltage (upper trace), which increases to match the signal.
low-sensitivity range 0 dB corresponded to about 50 watts on a steady sine-wave signal, or about 250 watts on brief musical peaks, into 8 ohms. The relative accuracy of the meter readings and range switches was good, and the least-visible meter deflection was at about 0.1 watt (sine wave).
------- FREQUENCY IN HZ (CYCLES PER SECOND) 100 200 500 CONTINUOUS AND EQUIVALENT SINE-WAVE WATTS/CHANNEL.
Comment. On the basis of its actual performance, the Soundcraftsmen MA5002 is a very powerful, clean amplifier capable of delivering considerably more power than many other comparable- or higher-price units. Not many $700 power amplifiers can deliver close to 1,000 clean watts to a pair of 4-ohm speakers (and not too many 4-ohm speakers can absorb 500 watts safely, so caution is advisable!). We found the protective system of the MA5002 to be excellent in respect to problems of shorted outputs or too many pairs of speakers connected in parallel.
In normal use, the MA5002 power amplifier indeed runs very cool, as one would expect from a "125-watt" amplifier whose output semiconductors (eight per channel) and power supply are designed for two to three times that power. If it is "pushed" it gets hot, of course, but we cannot imagine ever needing a fan or other external cooling means.
We listened to the amplifier at length, often driving it until all its signal lights except OVERLOAD were activated. At no time was there any audible evidence of anything out of the ordinary; the MA5002 behaved and sounded just like the very rugged and clean high-power amplifier that it is.
It is worth mentioning that, unlike some other techniques for increasing the short-term power-output capability of an amplifier, the Soundcraftsmen "Class H" system is capable of full-time operation. In other words, the extra power is available not only for a fraction of a second during a musical peak, but for as long as may be required by the musical pro gram material, without any damage to the amplifier or change in its operating characteristics. One may argue whether the Soundcraftsmen design truly represents a new class of amplifier operation, but it is certainly a very effective means of economically providing large amounts of clean audio power when it is needed.
Also see:
THE CASE FOR OPEN REEL--What it comes down to is technical specifications
Technical Talk, Julian D. Hirsch
NEW PRODUCTS--Roundup of the latest audio equipment and accessories (with vintage magazine ADs)
EQUIPMENT TEST REPORTS (Dec. 1977)
Source: Stereo Review (USA magazine)