|Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting
Departments | Features | ADs | Equipment | Music/Recordings | History
by Peter W. Mitchell
Americans buy about a billion dollars' worth of hi-fi components each year, and much of that hardware comes from the large manufacturers whose names everyone knows--Pioneer, Bose, Kenwood, Advent, etc. But one of the major trends in the audio industry is the proliferation of independent small companies, manufacturing a small number of products each of which represents a particularly thorough and thoughtful design effort by just one engineer or a small staff.
In the October and November issues of Audio, Bert Whyte described many of the exotic and innovative new products on display at last summer's CES, including some of the most advanced designs seen (and heard) at the Pick Congress. But there are so many companies designing new products these days, with so many exhibits both in the Pick and other hotels as well, that even the highly peripatetic Mr. Whyte was not able to cover them all, even in two months' worth of columns. So in the following paragraphs we will look at a few more of the provocative products seen last summer at the Pick; given the usual delays in starting up production of new designs, some of these components will just be arriving in retail stores when this issue gets to you. Being a dollar-conscious New England Yankee myself, I will concentrate mainly on products with non-exotic price tags, if only to show that a thoughtful and innovative de sign needn't cost as much as a new car. And rather than providing a laundry list of new products, I will focus on the ideas and concepts which stand behind them--new ideas worth knowing about and a few old ideas which deserve another look.
Discussions of new amplifiers often revolve around exotic technology--FETs, ring-emitter transistors, floating bias output circuits, slew rates measured in hundreds of volts per micro second, etc. But there is another school of amplifier design which suggests that while these improvements may be audible under idealized conditions, much larger differences in amplifier performance arise under typical operating conditions which are not considered in the design lab. By rigorously analyzing amplifier misbehavior modes and thoughtfully applying non-exotic technology, a surprisingly clean and musical amplifier can be produced at reasonable cost.
As a trivial example, an "ideal" amplifier might have vanishingly small levels of transient intermodulation or slew-induced distortions (TIM and SID), obtained by making the amplifier have very fast response from d.c. to television frequencies. But in typical real-world audio systems, all of the signal energy found at infrasonic and ultrasonic frequencies consists of distortion and interference whose amplification and delivery to the speaker yields no benefit and may produce audible intermodulation products within the audible range. By designing and placing a minimum-phase, audio bandpass filter ahead of the output stage, the designer can prevent a common real-world problem, eliminating a cause of unclean sound. At the same time this provides a gentle limit to the rise times in the input signal, thus ensuring that the output stage can never be driven beyond its slew-rate capability. The result is an amp which is free of TIM or SID without exotic and costly ultra-high-speed output devices.
A more serious set of real-world problems arises because amplifiers are designed to drive 8-ohm test resistors but are used to drive loudspeakers whose minimum impedance is usually lower than specified, partly reactive, and may become increasingly non-linear at high volume levels. In a simple experiment to measure the current flowing through loudspeakers while they were reproducing music, I have observed current peaks of plus and minus eight amperes (i.e. 16 amps peak-to-peak) in simple drumbeats played by America's best-selling "8 ohm" loudspeaker. With "4-ohm" speakers, or two 8-ohm speakers wired in parallel, this current requirement might have doubled. Very few of today's amplifiers, regardless of their exotic technology, are designed to provide such output currents to the speakers. Furthermore the impedance of a typical loudspeaker is partly reactive, i.e. inductive and capacitive as well as resistive, so currents and voltages tend not to be in phase. As a result, it is not uncommon to find an amplifier's protection circuit producing audible distortion at output levels lower than the amp's rated power. VI (voltage-current) limiters commonly work by feedback, altering the input signal in an attempt to keep the output within the prescribed "safe-area" bounds prescribed for the out put transistors.
The object of this prelude is to suggest that audible differences among amplifiers under typical operating conditions often depend as much on the amplifier's interface with the loudspeaker as on conventional performance parameters such as the distortion and slew rate measured when driving a resistor on the test bench. By minimizing or eliminating the influence of protection circuits on the amplifier's sound, a remarkably high level of performance may be obtained from an amplifier with non-exotic specifications. It is possible, for example, that the notably fine sound of the MOS-FET power amplifiers which have been introduced by Hitachi and Hafler is not due only to the high speed of the MOS-FET; it probably is also due to the MOS-FET's negative thermal co efficient and consequent resistance to self-destruction, leading to circuit de signs which require unusually little VI limiting for protection.
Two new amplifiers seen at the Pick Congress exemplify this design approach but use "conventional" bipolar transistors. One of these is from NAD, an international company which has a headquarters in London. NAD decides what products it wants to create and farms out their actual manufacture to specialized independent factories around the world (e.g., NAD amplifiers and tuners are assembled in Taiwan, their turntables are from Eng land, and a moving-coil cartridge comes from Japan). They had the most incongruous display at the show: A simple-looking little $175.00, 20-watt Model 3020 integrated amplifier driving the big Acoustic Research AR-9s. It didn't reach disco volume levels, of course, but its sound was clean, transparent, and detailed at levels substantially higher than what we usually expect from a low-cost budget amp. I am reminded that Paul Klipsch used to re mark that what the world needs is a good five-watt amplifier (to drive his high-efficiency Klipschorns); in our pursuit of the state of the art, we often neglect the needs of people whose bank accounts won't accommodate kilowatt amplifiers and whose neighbors won't tolerate wall-rattling volume levels. Stereo systems which produce only a modest quantity of sound need not be equally limited in quality. The NAD 3020's sonic performance probably can be credited to several factors--the use of large output transistors normally seen in 60-watt amplifiers, with voltage and current ratings sufficiently conservative that VI limiting never occurs in use; enough output current to drive even 2-ohm loads; a phono preamp which is unusually elaborate for a budget amp; a "soft clipping" circuit which minimizes power-supply buzz and even-order harmonics when the amplifier clips, and a high IHF dynamic headroom factor.
The second amplifier which exemplifies these ideas is the Model 1 power amp designed by Tomlinson Holman and his colleagues at Apt Corp. It uses newly developed bipolar output transistors with an exceptional combi nation of ruggedness and speed, yielding high slew rates and very high peak output currents without the limitations imposed by VI limiting. This power amp is nearly the same size as the well-known Apt/Holman preamp and is rated at 100 watts/channel; with a dynamic headroom of 3 dB, it actually exceeds 200 watts/channel on musical transients into either 8- or 4-ohm load impedances and maintains strong output down to 2 ohms. Its power transformer has two sets of secondary taps which are selected by a rear panel switch; depending on the true impedance of his loudspeakers, the listener can set the transformer to supply the large voltage swings required to maintain a large dynamic headroom with load impedances of 8-16 ohms, or set it to provide the lower voltages and larger currents required for high peak output into 2- to 4-ohm loads. Since listeners usually don't know the true impedance of their speakers, a VI comparator monitors the output voltage and current flow and activates a front-panel light to tell the user which way to set the transformer switch.
Apt also displayed a mockup of their Model 2 amplifier, a more exotic de sign which will be released next spring. It employs MOS-FETs (in multiples for large current capacity) and a Class-D switching power supply which eliminates the usual bulk of a large power transformer and filter capacitors. The Model 2, like the Model 1, is about the size of the Apt preamp and has a rated output of 200 watts/channel continuous and 800 watts/channel for transients! Normally MOS-FETs re quire a lot of heat-sinking because of their quiescent power dissipation (high idle current combined with the high power-supply voltages required or large rated power output), which would make the compact size of the Apt 2 impossible. So the Model 2 doesn't use fixed power-supply voltages; instead an active "smart" supply s used, producing supply voltages which vary over a broad range to track he audio signal. The compact size of he celebrated Carver "magnetic" amplifier is based on a similar principle forcing a supply voltage to track the audio input so that the quiescent dissipation is low at normal levels while a large voltage swing is generated on demand to accommodate loud passages and transients.
Currently all of the MOS-FET amps announced so far have names beginning with H: Hitachi, Hafler, Holman Apt), and H/H Electric of Cambridge, England. H/H has developed a line of amps employing fewer or more of the devices to yield rated outputs of 60, 150, or 200 watts/channel, with the option of bridged mono operation to yield 250, 500, or 800 W respectively.
My attention was drawn to the H/H amplifiers by the speakers that they were driving at the Pick: The Dayton-Wright XG-10 full-range electrostatics.
It is risky to judge loudspeakers on the basis of first impressions in a hotel-room show; any conclusions must per force be tentative. But while the earlier Dayton-Wrights had generally struck me as sounding a bit hard, my first impression was that the reproduction of both vocals and instrumentals through the XG-10 was startlingly realistic and uncanned. Such quality doesn't come cheap of course, $2,900.00 for a pair.
But that doesn't seem exorbitant by the standards of today's exotic loudspeakers.
Having delivered my disclaimer about trusting first impressions at shows, I must now echo Bert Whyte's enthusiastic comments about the remarkably uncolored, wide-range sound of the Bowers & Wilkins 801 ($2,500.00 per pair). In the carnival-like atmosphere of the show, the B & W rooms were a model of thoughtful attention to program material, room acoustics, and the aesthetic values of music. At any show one can always count on Allison Acoustics, too, for a similarly relaxed and civilized atmosphere with well-chosen music.
The principal trend visible in current speaker designs is the attention being paid to the control of reflected sounds that can bounce off the speaker's own cabinet or off nearby surfaces such as walls and floor. A decade ago at AR, Allison and Berkovitz discovered the irregular response that arises from the woofer's boundary reflections, and Allison Acoustics pioneered the increasingly popular column design with the woofer placed at the floor/wall inter section. But so far other manufacturers have not applied the same principles to smaller bookshelf speakers--in which the woofer ends up on the top or side of the cabinet, next to the wall, with the tweeter on the front. The new Allison Six is a particularly compact implementation of this approach, in the form of an 11-in. cube. Its spacious stereo sound field was not surprising, but its clean, deep bass output was--particularly when the Allison Electronic Subwoofer was used to dig into the recording and bring its bottom-octave bass up to audible levels.
There's a new speaker company formed every week or so, and one of the more promising is Boston Acoustics, founded by Andrew Petite and Frank Reed after they left Advent.
Their first speaker, the A200, is quite un-Adventish in appearance and sound; in fact, it doesn't look like any of the more traditional Cambridge bookshelf speakers. It's a floor-standing unit, only a few inches thick, 11 ft. wide and 3 1/2 ft. tall, with its 10-in. woofer near the bottom behaving as if it were mounted at the floor/wall intersection and radiating uniformly into that 90-degree solid angle with all boundary reflections in phase with the woofer's direct output. The midrange and tweeter, flush-mounted in a very large front-panel area and more than a half-wavelength away from any cabinet edges, radiate rather uniformly into a 180-degree solid angle. As a result the speakers have a remarkably uniform tonal quality over a large listening area.
A couple of really small speakers drew a lot of attention at the Pick: The KEF 101 minis mounted in midair on slim poles (like microphone stands), and the Cizek KA-1 minis beautifully finished in Hawaiian koa wood with a companion common-bass subwoofer also finished in the same wood--a striking solution to the perennial problem of integrating hi-fi technology into a gracefully appointed living room.
Historically there have been numerous obstacles to widespread acceptance of moving-coil pickups by American audiophiles--low output voltage (requiring the addition of a pre preamp or transformer), a typically massive cartridge body (implying low compliance and mediocre low-frequency tracking), a tipped-up high end with undamped ultrasonic ringing, styli that were not user-replace able, and a retail price substantially higher than the better moving-magnet designs. Happily these obstacles are gradually becoming overcome, though not all in any one model. For example, new MC pickups from Adcom and NAD have joined the family of high output-voltage models from Satin and Dynavector, feeding directly into the standard RIAA phono input without any step-up device (though, for the benefit of listeners who have already made that investment, the Adcoms are optionally available in low-voltage versions as well). Adcom's head, Newton Chanin, was previously an Ortofon executive, and the Adcom Crosscoils (so called because their coils are sym metrically wound on an X-shaped armature), will retail for around $200.00. Adcom personnel conducted a showdown at the Pick, an on-the-spot comparison versus many of the best-known MC pickups from Fidelity Research, Denon, etc.
The Adcom and NAD pickups are made in Japan, in part because a high output MC cartridge requires a large number of turns of extremely fine wire to form each coil--involving specialized techniques and equipment not widely available. The NAD 9000 costs only $160.00 and has a total weight of 6 grams, less in both respects than some of today's moving-magnet pickups. The NAD has an internal electrical damping network which eliminates ultrasonic ringing and yields an unusually flat response.
Meanwhile Satin has announced a new model in its popular 117 series, an MC pickup whose stylus assembly is user replaceable.
Yamaha and JVC have taken another approach to the problem of winding consistently good coils; instead of wire they employ tiny bits of IC substrate on which a delicate spiral conductive pattern has been photo-etched. One more aspect of cartridge design is receiving fresh attention: The cantilever bar which carries the stylus vibration up to the coil/magnet assembly. New MC pickups from Sony and Onkyo employ carbon fiber, Yamaha's MC-1 employs a beryllium rod, and Dynavector's Karat 100R has a solid ruby cantilever and a proportionally exotic price.
(adapted from Audio magazine, Dec. 1979 )
= = = =