A Comparative Survey of Power Amplifiers: Part 1 (March/April 1977, Vol. 1, No. 2)

Part I: In which we compare the sound of 10 power amplifiers for openers, mostly in the 100 to 200-watt class, and just begin to warm up to serious electronic testing.

We'll handle this a little bit differently from our preamp survey. There we presented what we could, or couldn't, find out in the laboratory as an introduction to our reports of the listening tests. Here we propose to let a series of sonically correlatable electronic tests emerge from our total experience with all the power amps as we conclude the survey, either in Part II or (if there's a spillover) in Part III.

The reason for this change of format is two-fold.

First, some of the most sophisticated units, with the greatest promise of providing new insights into electronic performance criteria, still remain to be tested as we go to press. For example, we're waiting for delivery of a Threshold 800A, which all sorts of experts assure us is The Ultimate. We've also been promised at least a brief exposure to a 100 watt-per-channel version of the Electro from Norway, of which the 25-watt version is as good as we've ever heard. Then there's the Bryston 4B, which we would have reviewed below if our sample hadn't come under suspicion for an elusive defect. (New sample coming; see also The Admonitor column in this issue.) We're also trying to get an Ampzilla II, and we've been promised a new pair of Futterman H-3aa mono units for more thorough examination. And so on.

Second, we have so far found that in the case of power amplifiers there really is some sort of correspondence between measurable and audible characteristics, and we're really hoping that as we gather more evidence we'll be able to assert this with greater authority. We have a feeling that, in this case, time is on our side and all we need is more data. Our preamp tests appear to hold out no comparable promise of maturation.

What we measured and what we didn't.

We had no reason to doubt the manufacturer's specifications in the case of any of the power amps we tested, so we didn't bother to verify them. They don't tell much about the sound of the amplifier in any case. The only exception worth noting at this juncture is the rise time, if at all specified, as it does seem to be related to listening quality (see under individual model headings, where applicable).

We therefore routinely check every amplifier that crosses our lab bench with square waves, both with resistive and with capacitive loading.

More about that at wrap-up time.

There are two offbeat tests, however, that we're very fond of and want to mention briefly even at this early stage of the game. One is a version of the CCIF intermodulation distortion test, with 14 kHz and 15 kHz mixed 1:1. It really separates the men from the boys. What's particularly interesting about it is that, in some amplifiers, it produces sidebands at 13 kHz and 16 kHz and in others a difference product at 1 kHz, but seldom both to any considerable degree. And each of these two kinds of IM sounds different. The other test is a measurement of what is formally known as network propagation delay and is more simply the time it takes for a signal to pass from input to output. It's performed on a dual-trace oscilloscope by subtracting the waveform of a pulse input from that of the output and examining the difference signal with delayed sweep. This is capable of yielding very useful information about the time-smear characteristics of the amplifier. We'll make brief reference to these and other simple tests under the individual reports while reserving judgment about their ultimate relevance until the conclusion of our survey.

The listening tests.

Our approach to listening evaluations was discussed in detail in Part 1 of the preamp survey and receives further attention in this issue in Part II of that survey as well as in the editorial article. We can therefore restrict our comments here to the specifics of the power amp tests.

Each amplifier was listened to through a pair of Dahlquist DQ-10's, modified with mylar capacitors in the crossover as described in our first issue. To those who don't particularly like the DQ-10 we can only say that you don't have to like it to use it as a tool; the fact that it's a difficult speaker to please, being totally unforgiving of anything but the cleanest electronics behind it, makes it very convenient for eliminating amplifiers that a/most sound right.

We don't consider the Dahlquist to be the ultimate challenge to a power amplifier; the "finalists" in Part II will all be listened to through full-range electrostatics, inefficient subwoofers that eat amplifiers for breakfast and other monsters. As a first hurdle, however, we don't know of anything more revealing than the DQ-10.

The preamp plugged into each unit under test was the revised Mark Levinson JC-2 with Lemo connectors, our current choice as a reference preamplifier. Since, so far, the lab tests on power amps seem to confirm the listening tests, we have no evidence of subtle interface or compatibility problems when using this pre amp with any of the power amps. The Rappaport PRE-1A with MC-1 head amp was also used in some of our tests to confirm the results.

The program sources consisted of records played mainly with the EMT cartridge but occasionally also with the Denon DL-103S and Grado Signature, plus a few second-generation master tapes played on a Revox A700.

We didn't find it necessary to run con trolled A-B listening tests on the power amps except in a very few, isolated cases. As will be apparent from the individual reports below, most of the units we tested had obvious short comings when listened to by themselves, and since we did find at least one excellent and relatively low-priced model, we didn't feel like investing additional time and effort into fine tuning our preferences among the not-quite satisfactory ones. We fervently hope that we'll have to A-B the bejeezus out of the few amplifiers that survive all of our test as we get into Part II. That will also be the time to analyze amplifier /speaker compatibilities-if indeed that's a realistic concept. (See the Quatre DG-250 review below on that subject.) With that we're ready to begin the individual reports.

Audio Research D-100

Audio Research Corporation, 2843 26th Avenue South, Minneapolis, MN 55406. Model D-100 High Definition Power Amplifier, 8995. Three-year warranty; not clear whether customer pays all freight. Tested #36401134, on loan by courtesy of Lyric Hi-Fi, White Plains, NY.

We have already made our comments anent Audio Research's venture into solid state; see the original SP-4 preamp review in our first issue and its follow-up in this issue. The D-100, a 100-watt-per-channel power amplifier, is the other vehicle in the incredible Analog Module promotion cooked up by ARC in an effort to beef up their State Of The Art image.

In our opinion, the only thing unquestionably SOTA about the D-100 is that it comes from Minnesota.

In all fairness, it must be pointed out that we were able to borrow our sample only for a very short time. Our exposure to the D-100 wasn't nearly as long as to some of the other amplifiers reviewed here; however, we're quite confident in our judgment that it isn't the best of the lot, either for the money or on an absolute basis. How high it rates as an also ran is another matter; possibly quite high.

The first thing that struck us about the D-100 when we started to listen to it was its completely sweet, unstrained sound, combined with a nice open quality. We've got a winner here, we thought.

But wait a minute. There was something wrong. How come we couldn't hear the precise overtone structure of the cymbals and tri angles? Wasn't there a lack of transient detail in general? There certainly was. What at first seemed sweet turned out to be merely blunted.

Then we also became aware of a lack of tightness and definition in the bass, all the way up to the lower midrange. This was no super amplifier, even if it was pleasant to listen to.

The output just wasn't a complete replica of the input.

Quick to the laboratory before we had to return the amplifier. The only other unit we had lying around there for reference was the GAS Son of Ampzilla (not quite our favorite, either). We looked at square waves passed through the D-100 and then the Son. The latter had a much shorter rise time, about half that of the D-100, which appeared seriously band width-limited by comparison. Eureka! Blunted square waves, blunted transient detail. It was almost too simple to be true, and we may yet decide that we jumped to a conclusion too quickly-but there it was. The lab appeared to confirm the ear.

We must add that the physical construction of the D-100 is every bit as impressive as the SP-4's. It's obviously built to last. Handsome, too, and very convenient to lift with those nice big handles. Still, we prefer to cut our fingers on the sharp corners of the handleless, slippery Quatre DG-250. It has better bass, better transient detail, more power, at half the price.

CM 912a

Audio International, Inc., 3 Cole Place, Danbury, CT 06810. CM 912a Stereo Power Amplifier, $899. Tested unnumbered sample, on loan from manufacturer.

This is from the former C/M Laboratories, now operating under a new name and a new management. Again, our sample was under our roof for a very short time only, not even long enough to be carefully looked at in the laboratory, but long enough to be listened to and ranked with considerable certainty in our sonic pecking order. It happens to rank quite high but not at the top. If none other, the Quatre DG-250 pecks it, and at a much lower price.

The CM 912a is a 150-watt-per-channel unit (225 at 4 ohms) with peak-reading meters and incorporates circuitry to drive the CM servo-feedback speaker, a feature on which we have no opinion. In conventional hookup, it sounds bright and detailed, some would say marginally (but not obnoxiously) over-bright, others would say just right. We definitely prefer the slightly more reticent but somehow more transparent, more natural highs of the Quatre. The bass of the CM 912a is very accept able when listened to by itself, but switching to the Quatre gives an immediate impression of greater tightness, detail and reserve power, despite the lower power rating.

An interesting and very elusive anomaly of the CM 912a is a kind of ping-pong or double-mono effect in stereo imaging. Plenty of left and right information, and strangely little fill-in in the middle. Whether this is some peculiar phase problem or merely an artifact of our equipment chain (or even just a figment of our imagination), we didn't have enough time to determine. We didn't hear it when we tested other amplifiers.

Overall, the CM 912a is a power amplifier we could probably live with quite happily if we didn't know what we know about others.

Electrocompaniet

Electrocompaniet, Tovengt. 14, Oslo 1, Norway. 'The Two-Channel Audio Power Amplifier,' price not avail able. Tested #86, on loan from private owner.

Okay, audio freaks, eat your hearts out.

Here's what we think is the world's best sounding power amplifier and (a) you can't buy one in this country and (b) it's much too low powered to be practical.

Electrocompaniet is simply the Norwegian form of The Electro Company, and we've been referring to this little marvel as the "Electro from Norway" (not to be confused with the Electro Research from California, which we haven't tested yet). The power output of this unit is approximately 25 watts per channel (we believe it's rated at 28 watts but it doesn't quite seem to make it); there's also a 100-watt per-channel version, we're told, but just try to get one. (We're trying. A friend who had personal dealings with the company in Oslo and ordered one directly from the factory many months ago is still waiting for it.) The sound of the 25-watt Electro requires only a brief review. It's the best. Of course, you can't push the amplifier beyond its power capability. As long as you play it about 10 dB below window-rattling level, you'll hear definition and transparency a whole order of magnitude better than with the Quatre, which is our top choice among commercially avail able units so far. The Electro makes the bass cleaner and tighter on the naked Dahlquist DQ-10 than it is normally with a biamped subwoofer. The midrange is pellucid, and high frequency transients are completely etched and focused, without the slightest ringing or fizziness. Of course, there comes the moment when you've just got to have more sock, and the Electro doesn't have it. Frustrating as hell.

We understand that the circuit has the blessing of Matti Otala, the Finnish archenemy of TIM, although it's incorrect to refer to it as "the Otala amplifier," as some people have, since Otala isn't in the business of designing circuits for the audio industry.

Whatever its provenance, the circuit is certainly fast and has tremendous bandwidth.

It just barely rounds the corners of 100 kHz square waves. We weren't allowed to have the 25-watt unit long enough to tell you more than that; if we ever get our hands on the 100-watter we'll certainly try to wring it out in the lab. Meanwhile, we're just sitting here savoring the memory.

Futterman H-3aa (preview)

Futterman Electronics Lab, 200 West 72nd Street, New York, NY 10023. H-3aa vacuum-tube power amplifier (mono), 8260; stereo pair, 3520. Auditioned manufacturer's demo samples.

Editor's Note: This isn't really a test report, since the equipment was made available to us only for a few hours. But we did have a chance to insert it into our reference system and form a rather firm opinion of its relative merit.

The Futterman output-transformerless vacuum-tube power amplifier has been an audio cult item for the past twenty years or so, and it's getting better every year. We had lived with at least three versions of it before being ex posed to this latest incarnation. An outfit in Michigan called Golden Ear now owns the patent to the circuit, but we haven't seen any amplifiers coming out of there; Julius Futterman, the original designer, is permitted under the patent-sale agreement to build a small number of hand-wired units per year, and it was a pair of these that we had the opportunity to audition.

The H-3aa is strictly a utility model but extremely well built; each mono chassis has six beam pentodes in its output circuit and can deliver 100 watts at 8 ohms (150 watts at 16 ohms-it increases the opposite way from transistor circuits). Coupling to the speaker is via 2500-microfarad capacitors, so that with an 8-ohm load the low-frequency response is down 3 dB at 8 Hz (by definition) and about 1 dB at 16 Hz. With the Dahlquist DQ-10, which has somewhat lower impedance than 8 ohms at the lower frequencies, these figures are a little bit less good, which still doesn't explain why we heard a slight looseness in the bass, since the damping factor of the Futterman is extremely high. In fact, this small deficiency (which, ac cording to Julius Futterman, can be remedied with even larger output capacitors) was the only audible flaw of the H-3aa during our brief exposure to it. In every other way, it sounded magnificent.

We were especially impressed with the highs, which were truly sweet and free from strain, while still extremely detailed. Only the Electro from Norway was comparable (or superior?) in this respect; how we wish we could have had both of them available at the same time for A-B-ing! The midrange of the Futterman was also superb; it couldn't have been more transparent, or more accurate in depth perspective. Since the H-3aa has ample power for all but the most inefficient speaker systems, it would undoubtedly be our reference amplifier if the bass could be tightened up just a wee bit. We shall see; we've been promised a pair in time for Part II of this survey.

On top of everything else, the price is extremely reasonable; you should really order a pair and see how much further improved they'll be in 1984, which is approximately when you can expect delivery.

GAS Son of Ampzilla

The Great American Sound Co., Inc., 20940 Lassen Street, Chatsworth, CA 91311. Son of Ampzilla stereo power amplifier, $434 (with rack-mount panel and handles). Five-year warranty; customer pays all freight. Tested #400845, owned by The Audio Critic.

Even though we cringe every time we pronounce its embarrassing name, we consider the Son of Ampzilla to be an excellent little power amplifier-not even so little, since its 80-watt per-channel rating at 8 ohms jumps to 150 watts at 4 ohms and even more at 2 ohms, making it especially useful for multiple-speaker and other low-impedance applications.

If the Quatre DG-250 didn't cost only a few dollars more, the Son would be a serious contender for best-sound-per-dollar rating in this survey. But we don't find the Son to be quite as accurate a reproducer as the Quatre, either audibly or measurably.

The sound of the Son is typically GAS eous; highly listenable, never harsh or distressful, with a neat trade-off between transient detail and rounded pleasantness. It's neither ultra-transparent nor veiled, just a little bit loose and whompy in the lower part of the spectrum. Switching to the Yamaha B-2, for example (just to give you an idea), results in a spectacular improvement in openness and freedom from whomp, but considerably harder highs.

In the laboratory, the Son shows outstandingly good square waves with a resistive load but rings rather badly when an additional 1-microfarad capacitor is connected across it.

This happens at frequencies as low as 2 kHz.

At 20 kHz the ringing becomes disastrous; the square wave is barely recognizable. This augurs ill for driving electrostatics, despite the amplifier's unusual capability to handle low impedance loads; indeed, a brief exposure to the sound of the Son through the latest Dayton Wright left us very unimpressed. Phase shift is quite high at 20 kHz and higher than that of the Quatre DG-250 even at 2 kHz; on the other hand, propagation delay is extremely low, which just may be what keeps it all from sounding opaque and '"'electronic." It's the CCIF test for IM distortion, however, that really nails this amplifier. With 14 kHz and 15 kHz going in (mixed 1:1), there's too much 13 kHz coming out: 0.2% at full power and more than 0.1% even at around 20 watts. The 16 kHz sideband is also there, but very little of it; more significant is the rarely seen | kHz difference product of almost

0.1% at full power and more than 0.06% at 20 watts, which may explain the whomping, since it indicates that fairly low-frequency garbage is being dumped by the more subtle high-frequency interactions. We can't really prove that any of this stuff is audible, but then how come the Quatre has so much less of it and sounds better? In view of this company's advertising and our experience with their Thaedra/Thoebe preamp design, we come to more or less the same conclusion about the Son of Ampzilla: A very good piece of equipment but not good enough to make you go ape.

Luxman M-4000

Lux Audio of America, Ltd., 200 Aerial Way, Syosset, NY 11791. Model M-4000 Power Amplifier, $1495. Three-year warranty, manufacturer pays all freight. Tested #16101103, owned by The Audio Critic.

Like several other items in the Luxman line, the M-4000 is beautifully made, highly luxurious, very expensive and not quite good enough.

What we mean is best illustrated by one experience we had with it when driving the Duntech DL-15 speakers. As we reported in our first issue, we're quite impressed by these speakers (now apparently off the market), al though they aren't our top choice. Our Associate Editor liked them, too, despite some flaws he pointed out the first time he heard them. The second time, several days later, he said, "You know, I was wrong about these speakers. I don't like them at all any more. They're harsh and nasal." Then we re membered. We disconnected the Luxman M-4000 from the Duntechs and substituted the GAS Son of Ampzilla, which was how he had originally heard them. "Ah," he said, "now I like them again." Yes, the M-4000 has just a touch of transistory hardness, nasality and a not-quite open quality. Nor is its deep bass the most detailed. It's still a lot better in all these respects than many amplifiers, but then we don't give consolation prizes. Its 180-watt-per channel rating is quite aan giving it good reserve power on piano reproduction for example, which seems to be its forte.

We also like its combination of VU meters and LED peak-reading display, which gives you the best of both worlds (like wearing suspenders along with your belt). Nicest of all are the input attenuators, calibrated in 1-dB clicks; what a gorgeous pair of controls and how we wish every good power amp had them! But, let's face it, sound comes first.

Again, it's the 14-plus-15-kHz IM test that sinks this luxury liner. At only 72 watts, the 13 kHz sideband alone amounted to 0.25% in one channel and 0.2% in the other. Taking the power down to 35 watts barely resulted in an improvement. The 16 kHz and 1 kHz IM prod ucts were more acceptable (especially the 1 kHz), but that much 13 kHz is almost certainly audible and probably accounts for the harsh ness. As they say in the commercials, CCIF sure works for us.

The M-4000 also showed some interesting anomalies with pulse testing, but before we could analyze the results, we had a chance to sell it (this one we owned). We grabbed the money and ran like a thief.

Quad 405

Acoustical Manufacturing Co. Ltd., Huntingdon, Combs. PEIS 7DB, England. Quad 405 Current Dumping Power Amplifier, $410. One-year warranty, customer pays all freight. Tested #2311, owned by The Audio Critic.

"Current dumping" is one of the few thoroughly original amplifier design concepts to come on the scene since the dawn of the hi-fi era. Its circuit details are beyond the scope of this review, but the idea, simplistically stated, is that one part of the amplifier supplies the brute force and a separate part fills in the fine detail. This is supposed to make the fine detail more completely realizable and the entire operation of the amplifier more stable. There are no adjustments (bias, etc.) inside the box (which, incidentally, is small, handsome and beautifully finished); the amplifier is set for life and replacement parts, if needed, are simply soldered in.

In all fairness, however, we must report that a very knowledgeable circuit designer we know has been heard referring to the 405 as "the Quad shit-dumping amplifier." This scares us because, as we said, he's very clever; maybe he knows something we don't. We happen to like the Quad 405 a lot; again, if the Quatre DG-250 weren't almost as reasonably priced, we might consider the 405 as a candidate for best value in our survey.

We especially like the bass; despite the low-frequency input filter (down 1 dB at 20 Hz and 12 dB at 7 Hz), the lows are firm, detailed and authoritative. The highs are, if anything, mildly subdued; at least there's not a trace of harshness or fizziness. The midrange is quite nice and open. Switching to the Quatre reveals, however, that there's a long way to go in all these departments; the DG-250 has tighter bass, a more transparent midrange, more crystalline highs. More power, too; the Quad 405 delivers only 100 watts into 8 ohms as against the Quatre's minimum of 125, and at 4 ohms the Quad is down considerably whereas the Quatre is up.

One must remember, of course, that the 405 was designed with the Quad electro static as reference speaker; in fact, it has provisions for a pair of plug-in limiter. resistors to protect that speaker from an output voltage of more than 20 volts. We can't tell you at this point whether there are any amplifiers that drive the Quad electrostatic even better than the Quad 405; the two certainly make beautiful music together. (As you probably know, it's inadvisable to connect just any old amplifier to the Quad speaker.) One more thing. If you hear hum and noise through the 405, it's probably your preamp.

The amplifier has an input sensitivity of 0.5 volts for full output, and that's a lot of gain.

The best thing is to use the low-gain Quad 33 preamp with it; as we've said before, the 33 plus the 405 are still quite a package for $675, even if they don't beat everything in sight.

Besides, if you don't give Quad and the pound sterling your continued support, they'll never come out with the new full-range electrostatic they're supposed to have under wraps there in Huntingdon, which does beat everything in sight-or so we're told.

Quatre DG-250

Quatre, 8223 Remmet Avenue, Canoga Park, CA 91304. DG-250 Gain Cell power amplifier, $495. No warranty information enclosed with original factory container. Tested #473, owned by The Audio Critic.

Before anything else, let's make it clear that this is a new Quatre amplifier, called the DG-250 Gain Cell, available only as of late 1976 and not at all like the previous Quatre power amp. The DG-250 uses a highly original analog multiplier circuit (not to be confused with Audio Research's so-called Analog Module, which in our opinion is just a pro motional name for IC's). The Quatre circuit is based on the philosophy of "if you can't lick em, join 'em." You can't lick the inherently nonlinear (logarithmic) output current characteristic of transistors, so you just accept it and put the signal through an oppositely non linear (antilogarithmic) process. The net result is linear-without feedback and hence without TIM. (They'll never teach it this way at EE school, but then the tuition there is more than $28 a year.) Actually, the DG-250 does have some feedback around the output stage, but it's still quite different from the typical amplifier that relies heavily on feedback to keep it linear from input to output.

Whatever the reason, the Quatre Gain Cell sounds awfully good. Of the power amplifiers we have evaluated so far, only the Electro from Norway sounded decidedly better from top to bottom (but only within its very limited power capability), and the Futterman H-3aa on top and in the midrange but not in the bass.

Overall, we hear very little that's wrong with the DG-250. The bass is extremely firm and tightly controlled, the midrange open and quite accurate in depth information, the highs clean and neither subdued nor exaggerated. What's more, the apparent reserve power is greater than you'd expect from the official 125-watts per-channel rating. The available power is probably closer to 200 watts, especially at lower impedance. (Our Dahlquist DQ-10 is in the 6-to-7-ohm range at the lower frequencies.) The measurements, lo and behold, bear out the sound. It's almost too neat to be true.

Negligible IM distortion with the 14-plus-15 kHz killer test. No propagation delay to speak of up to 50 kHz. Quite a bit of ringing on 20 kHz square waves into a load of 1 microfarad across 8 ohms, but the competition is even worse in that respect.

As a matter of fact, we're beginning to think that the Quatre and the Dahlquist don't form a synergistic combination, as we suggested in our first issue. Maybe the Quatre is simply a better amplifier, period. Maybe compatibility is just a name for our ignorance of the more subtle factors of electrical performance. How about that, golden-ear innocents? By the way, the Quatre Gain Cell looks exactly as the name suggests. A big, black, anodized aluminum brick without anything sticking out of it. There's a tiny red LED in front and some heat sinks in the back. The edges and corners are sharp. Something like that monolith in 2001: A Space Odyssey.

In fact, we can just see the apes from the GAS ads and posters crowding around it and reaching out in awe to touch it. For only $61 more than the Son, the dawn of a new era. . . ? SAE 2400L Scientific Audio Electronics, Inc., PO Box 60271, Terminal Annex, Los Angeles, CA 90060. 2400L Solid State Stereo Power Amplifier, $800. Five-year warranty, not clear who pays freight. Tested #24-03019, on loan from rep.

This is a 200-watt-per-channel power amp with fully complementary circuitry from input

to output (not every amplifier in the SAE line is as up-to-date in concept) and a very neat LED peak-reading power level display. The front panel is of the professional rack-mount type, with big handles. But the sound is merely good, not outstanding.

We liked the bass best; it's quite firm and authoritative. The highs are less hard and grainy than we've heard out of a lot of expensive amplifiers, but far from perfect. There's still a residual amount of aggression. The midrange is rather closed down; indeed, there's a dark, syrupy quality to the overall sound, which at first passes for smoothness but is soon perceived as opacity and lack of detail.

Listenable but not accurate.

The laboratory observations are revealing.

There's no high-frequency harmonic distortion worth mentioning and rather low distortion with the CCIF test. But we measured considerable propagation delay, as well as phase shift, throughout the audio range. Both were quite evident even at frequencies as low as 2 kHz.

Doesn't that suggest a correlation between measurable time smear and audible blurring of detail? Furthermore, with square waves from 5 kHz to 20 kHz there was appreciable ringing even into a purely resistive load. That's unusual and may account for the subtly un comfortable sound at the higher frequencies.

All in all, the 2400L is just another amplifier. Neither its virtues nor its faults are interesting enough to make a fuss over.

If at least they had apes in their advertising . . .

Yamaha B-2

Yamaha International Corp., PO Box 6600, Buena Park, CA 90620. NS Series B-2 Stereo, Power Amplifier, 8850. No warranty information enclosed with original factory container. Tested #02834, owned by The Audio Critic.

Several references to this 100-watt-per channel vertical- FET power amp in our first issue (Part I of the preamp survey and else where) indicated that we rate it quite high, almost as high as the Quatre DG-250. We're not so sure anymore.

The B-2 undoubtedly has a strikingly open sound, so open that switching to most other amplifiers creates the impression that some thing has closed down. This quality is hard to resist, but that isn't all that's hard. So are the highs of the amplifier; in fact they're downright brittle and unpleasant. The seduction of the openness is gradually replaced by irritation as one listens further. We're now inclined to believe that what we thought was special compatibility between the Mark Levinson JC-2 and the Yamaha B-2 is simply a case of reduced irritation, the explanation being that the JC-2 with its super-clean highs makes no contribution to the total distortion, whereas other preamps do. (Slew-rate-related compatibilities are an 42 other matter; the B-2 has an unusually high slew rate of 60 V/uS, which is a subject we won't go into until Part II of this survey.) Since, obviously, a great deal of engineering sophistication has gone into the circuit design of this true DC amplifier, we don't want to be hasty in our dismissal of its sonic performance, but (again!) the audible problems at the higher frequencies are reflected in the results of the CCIF test.

At approximately half power (50 watts), 14 kHz and 15 kHz mixed 1:1 throw a fairly large sideband at 13 kHz (0.22%) and a smaller one at 16 kHz (0.06%)-about as bad as the Luxman M-4000 and worse than the GAS Son of Ampzilla. On the other hand, there's no 1 kHz product, so the net correlation is over brightness but no whomping. (Too simple? Maybe, but it's plausible.) The openness and clarity seem to be borne out by very low propagation delay (almost as low as in the Quatre); what there is of it is constant at all frequencies. All very neatly accounted for, isn't it? One can't conclude a review of the Yamaha B-2 without expressing admiration for the marvelous peak-reading meters and the two pairs of inputs, making it possible to A-B two preamps or other sources at the touch of a push button, with instant monitoring of levels. We'd love to make the B-2 our reference amplifier just for that reason, but we couldn't live with those highs in the long run.

Somebody who is very knowledgeable about these things told us recently that what seems to be bothering us is the typical East Coast sound. From the East Coast of Honshu.

Recommendations Since neither the Electro from Norway nor the Futterman H-3aa really "exists" from the consumer's point of view (just try and get one), it becomes a very simple task to designate our top choice-so far. Don't forget, though, that some very sophisticated units are still waiting to be tested.

Best power amplifier so far, regardless of price: Quatre DG-250 Gain Cell.

Best sound per dollar: Quatre DG-250 Gain Cell.

-------

-------

[adapted from TAC, Vol.1, No.2]

---------

Also see:

Fishing for Bass: A Look at the Subwoofer Scene

Various audio and high-fidelity magazines