Preamplifiers Revisited (No, Not Again!) (Vol.1, No.5: Winter 1977/78)

Yes, again. Because the situation has changed considerably, and for the better. The unlistenable horrors are slowly fading into the past, close-to-SOTA performance has become very affordable (would you believe $3797), and the best has gone right through the roof.

Everybody is making a preamplifier these days; it seems to be the in thing to do, or at least the preferred way of obtaining entree into the audiophile marketplace. The number of pre amp-only manufacturers, and of those who started out that way, is legion. The consumer is, of course, the beneficiary; good preamps are becoming easier to find than bad ones, and the price of the good stuff appears to be trending counter to inflation even if not exactly plummeting.

Comparing this latest batch of preamps to the ones covered in our Number 1 and Number 2 issues, we discern a distinct improvement; even the worst of them is considerably more listenable than, say, the BGW 202 (now apparently off the market).

We're still having considerable difficulties trying to correlate our laboratory measurements of preamplifiers with what we actually hear coming out of them through our reference system. The new measurement procedures we're about to phase in for our power-amp evaluations will be at least partly applicable to low level amplification stages as well (see the power amplifier article in this issue, so that we won't have to repeat ourselves needlessly here); we expect these new methods to give us a better understanding of the measurability/audibility relationship in preamp distortions, but they aren't reflected yet in the reviews below. Even so, the tests reported here differed to some extent from the earlier series, owing to certain refinements of our usual measurements and listening procedures, so that a few comments are in order.

The RIAA equalization mess.

We now have instrumentation for measuring the RIAA equalization of preamplifiers very quickly, completely and accurately. (It used to be a slow and painful chore.) As a result, we've become aware of some interesting anomalies, goofs and impasses.

First of all, there's complete anarchy among preamps at the top end of the audio range. If the 20 kHz response is to be exactly -19.6 dB relative to 0 dB at 1 kHz, as the official playback characteristic dictates, then the 6 dB per octave slope must be continued to a very high frequency in the equalization network. If the circuit designer finds it necessary to stop the roll-off with a "corner" at, for example, 50 kHz, there will be an error of a fraction of a dB at 20 kHz and at even 15 kHz.

Such corners are put in at different frequencies by different preamp designers; furthermore, they're also put in at different frequencies at the cutter end, in the recording preemphasis curve. (We're told that the Neumann and Ortofon cutter head people have finally gotten together on this and will henceforth provide identically leveled-off preemphasis curves, but what about your present record collection?) We strongly suspect that worst-case mis matches in this preemphasis/roll-off process will be audible on fine-tuned systems.

The solution would be an emendation of the standard; in addition to the specified time constants of 3180, 318 and 75 microseconds, a fourth one (say, 3.18 microseconds) could be included, so that the high-frequency slope doesn't by implication extend to infinity. But what do the standard committees come up with instead? Nothing on the high-frequency end- but a needless and unproductive complication on the low-frequency end.

A new IEC Recommendation, which has not yet been made a standard by the RIAA but is already being used by a number of preamp manufacturers (for example, Audionics and Hafler), adds a 6 dB per octave bottom end roll-off to the existing playback curve, with the -3 dB inflection point at 20 Hz. This is ostensibly to keep undesirable low-frequency energy (from record warps, turntable rumble, arm resonances, etc.) out of the power amplifier and speaker, but it happens to be more effective as an obligatory tone control for bass cut than as a subsonic filter. At 5 Hz, the low-frequency garbage (if it's there) is attenuated only by 12 dB, which isn't enough; on the other hand, the response is already down 1 dB at 40 Hz, which is audible on a system with good bass response. Since standard committees are hard to argue with after the fact, the best way out under the circumstances would be to make this extra RC roll-off available via a separate two way switch (marked RIAA old/RIAA new), so that the preamp would be officially up-to date but still not irretrievably frozen in the new format. One thing is certain: no record will be cut with the inverse of this new play back curve, as it would require ridiculously wide groove spacing.

Another widespread RIAA boo-boo is in the equalization network used for creating the required time constants. Too many circuit designers choose a topology that can't possibly work accurately, even with resistors and capacitors of the tightest available tolerance.

The circuit elements interact in such a way as to produce more than just the intended inflection points; there are additional little bulges and saddles in the response that simply can't be trimmed out, only moved from one place to another. The likelihood of audible colorations depends on the location and amplitude of these bulges and saddles. The whole problem is predictable and avoidable but too often not predicted and avoided.

Listening test conditions.

As we stated the first time the subject came up, the only justifiable purpose of an expensive preamplifier is to reproduce phonograph records accurately. (If all you need is a source selector switch and a volume control, there are simpler and less costly solutions.) Therefore the only way to evaluate the listening quality of a preamp is to play records through it-with the best possible cartridge and arm.

Since moving-coil cartridges are far ahead of conventional magnetics in bandwidth, slewing speed, dynamic range and freedom from time smear, we're now convinced that the only way to wring out the phono stage of a preamp to the limit of its capability is to drive it with the best MC one can find. In our case that was the GAS 'Sleeping Beauty' Shibata (mounted in the Breuer Dynamic 5A tone arm on the Thorens TD 126 Mk IIB turntable-see the reviews in this issue). The impedance match to the phono input was effected via the Verion Mark 1 transformer, the only completely neutral, no-sound-of-its-own device for that purpose known to us. The approximately 80 kHz bandwidth of the Verion allowed most of the out-of-band program components coming out of the cartridge terminals to get through to the first stage of the phono circuit and separate the preamps that could handle that kind of input from those that couldn't. Re member, even with perfect lateral and vertical alignment, and even with an optimally tracing Shibata tip, there's more than just audio coming out of those grooves; a fast and ac curate transducer such as the 'Sleeping Beauty' will generate fairly high-amplitude out-of-band signals, which will either pass through without a hitch or excite in-band IM products, depending on the phono circuit design.

This is one of the important things that Mark Davis ignored in his experiments “proving" that all preamps sound alike; the Shure MOII1E he used is a slow, bandwidth-limited device that makes very light demands on the phono input stage. You might as well test drive an automobile at 30 miles an hour and then decide that it handles just like all others.

For the same reason, we mistrust those straight-wire bypass tests of preamplifiers that use a tape recorder through an inverse RIAA network as the program source. Quite aside from the fact that it's never 30-ips wide-track master tape that's used (so as to get some bandwidth and dynamic range in there, comparable to what you can get off a direct-to-disc recording with an MC cartridge), these tests also fail to evaluate out-of-band program component handling. No tape recorder we're aware of generates the out-of-band spectrum of a fast MC cartridge. Mind you, it isn't just tracing anomalies that create that spectrum. The generation of FM sidebands is inherent in the cutting/playback process itself.

All of these criteria are, of course, largely academic when the alignment for optimum lateral and vertical tracking geometry is neglected. It just isn't possible to listen critically to a preamplifier with the tone arm and cartridge incorrectly aligned. Much as we like to compare our results with those of other practitioners (including our subscribers), we really feel imposed upon when challenged to justify our findings to non-aligners with contrary opinions (e.g., Mark Davis). We don't believe the issue is negotiable.

Thanks to our expanded facilities, we were able to place all the preamplifiers reviewed below on open shelves at the same time and switch back and forth among them at will. All channels were adjusted for exactly the same gain at 1 kHz, from the cartridge terminals to the power amp input, within no more than 0.25 dB. At first we used an extremely costly "straight-wire" switching system for our listening comparisons and wasted an unconscionable amount of time. The switcher turned out to have subtle (but eventually measurable) short comings that put a barely perceptible haze over the sound and tended to homogenize the differences among the preamps. We then changed to A-B-ing by a lightning-fast manual plugging and-unplugging technique that necessity forced us to master, and our conclusions are based on this cumbersome but far cleaner modus operandi. (Incidentally, we wonder about the switching system in the Mark Davis experiments. The design of an unquestionably straight-wire switcher is more than just a routine engineering problem.) The reference speaker/amplifier system used in these listening tests was the Beveridge 2SW. With all its minor but frustrating flaws, it lays down a more lifelike sound field and resolves inner detail far more accurately than anything else known to us. There was really no other choice.

Do they really all sound different? About half of the preamps tested followed the RIAA equalization curve so accurately that they tracked one another in frequency response within better than 0.2 dB. Phono input capacitance as a potential source of frequency response variations through cartridge inductance interaction was washed out by the 6-ohm MC cartridge used (even through the transformer).

Volume levels, as we said, were accurately matched. So the basic requirements of the Mark Davis they-all-sound-alike school were satisfied in many, if not all, of our listening tests. And-we heard differences. In some cases small and subtle ones, in other cases laugh ably big ones. Therefore we now reaffirm our original conclusion that Mark Davis and the Boston Audio Society egalitarians are wrong.

(Are-oh-en-gee, as one of our friends likes to emphasize it. But see also the lengthy correspondence on the subject in the Box 392 column of our last issue.) That said, we must admit that in certain cases two preamplifiers of different topologies can sound remarkably alike, even if not necessarily identical. We suspect, for example, that an open loop of limited bandwidth plus gobs of feedback to widen the response and reduce static distortions to triple-oh figures is a great "leveler" of sonic differences. (See also the power amplifier article in this issue.) Tracking these small differences in a blind A-B test re quires extreme concentration that some auditioners find unusually fatiguing, so that they develop an I-give-up attitude even though they discriminated fairly accurately on their first few tries. Others manage to latch on to the sonic signature of either A or B or both and never make a mistake. We must emphasize that the difficult cases constitute the exception rather than the rule; anyone, for example, who can't immediately hear the difference between the latest Mark Levinson ML-1 and the comparably dead-flat GAS 'Thalia' should immediately consult a qualified otolaryngologist.

On to the reviews, then; and remember that as our test procedures become progressively more positive and revealing, there's less and less need for circumlocutory audio-word painting, so that the how-liquid-is-the-upper midrange crowd may not be getting their money's worth here, even though they'll find out which preamp to buy and why.

Ace 3100

Ace Audio Co., 532 Fifth Street, East Northport, NY 11731. Model 3100 Stereo Preamplifier, 3325 (with external power supply). Two-year warranty. Tested #P4016, on loan from manufacturer.

This rather Spartan but definitely audiophile-oriented unit uses the 4739 op amp chip in the phono stage. Since the incredibly successful Hegeman circuit module is built around the uA739, which is presumably the same thing (or almost the same thing), it's interesting to observe that the Ace 3100 also shares the Hegeman preamp's marvelously open, focused and detailed midrange. It's right up there in the near-SOTA category in that respect, but all is for nought on account of the highs. They're sizzly and nasty and well-nigh unlistenable.

What a shame; this preamp almost makes it to the top at $325.

We don't think the RIAA equalization is the culprit, although its error curve is typical of the two-humps-and-a-saddle profile we discussed above. We'd call it +0.8 dB, with one of the humps centering on 7 kHz or so, but that alone wouldn't make the highs sizzly. We measured 20 kHz harmonic distortion in the high single-oh region at two volts out, most of it coming from the line amp, not the phono stage. We don't think that explains the funny highs, either.

We'll just have to let Ace figure out what went wrong.

AGI Model 511A

Audio General, Inc., 1631 Easton Road, Willow Grove, PA 19090. Model 511A Stereo Preamplifier, $465. Three year warranty, manufacturer pays all freight. Tested #7410167, on loan from manufacturer.

The "A" revision of this beautifully constructed and, from that point of view, very reasonably priced preamp differs in minor particulars from the original 511, so we retested it. Ours was the basic version having 33 dB of gain in the phono stage rather than Option H, which provides 40 dB of gain with the idea of sparing you the need for a transformer or head amp when using the higher-output moving-coil cartridges (not a low-noise solution, to say the least).

In our current reference system, the S11A exhibited a somewhat edgy, irritating quality that we don't recall in the 511, which we had considered quite smooth. Of course, the GAS 'Sleeping Beauty' Shibata cartridge through the Verion transformer will excite whatever transient problems exist in the input circuit.

We heard a tiny burst of harshness every time there was a dynamic peak in the music. Also, there was an overall effect of thinness-or perhaps a lack of adequate fullness-in the sound, even though the RIAA equalization of the 511A is not far from perfect.

Don't misinterpret the above comments, though; this is still a very decent preamp, roughly of the same order of quality as the D B (although we tend to lean toward the latter in our preference). On the other hand, it isn't quite in a class with the Van Alstine, which in turn is still far from the winner's circle.

What's more, even if we happen to be wrong about these fine-tuned rankings, the Audionics (see below) is so clearly the top choice in the $400-ish range that the rest is academic.

Audionics BT-2

Audionics, Inc., Suite 160, 10950 SW 5th, Beaverton, OR 97005. BT-2 Preamplifier, $379 (with handles on front panel, $399). Three-year warranty. Tested #02350, on loan from manufacturer.

Here's the surprise and delight of the new preamp generation: a $379 unit that turned out to be one of the top three or four we had ever tested, at any price. Audiophiles with big ideas and limited budgets, rejoice! The only preamps we know of that we clearly prefer to the Audionics BT-2 are the Hegeman HPR /CU (8645) and, one small step up from there, the latest Mark Levinson ML-1 ($1250). That's all. The latest Rappaport PRE-1A ($755) has smoother highs and per haps a wee bit better inner detail, but it's also thicker and more nasal in the midrange, where the Audionics excels. Others don't even come into the picture, meaning that the Audionics stands alone in its price range and then some.

Overall, we'd characterize the sound of the BT-2 as very open, smooth, beautifully defined, and extremely listenable. It lacks only the startling immediacy and super detail of our top choices.

In the laboratory, we found nothing important to criticize except the way the "new" RIAA playback characteristic is handled.

From 1 kHz on up, the error is +0.0 dB all the way up to 30 kHz (wow!), but going down ward it rises from -0.3 dB at 200 Hz to +0.4 dB at 39 Hz and then dips to -1.3 dB at 20 Hz. This is a ripple instead of a first-order roll-off, corresponding to neither the new nor the old standard. Strangest of all, there's a "filter defeat" switch (exactly according to our recommendation as discussed above) that increases the amplitude of the ripple by several dB in stead of straightening out the response! So, all you bass purists, keep your cotton-pickin' hands off that defeat switch and leave it in "normal." (P.S. With the filter in, we heard no bass anomalies.) In any event, we won't let this small blunder dampen our enthusiasm for this outstanding "best buy". Well done, Audionics.

DB Systems DB-1/DB-2 (follow-up)

DB Systems, PO Box 187, Jaffrey Center, NH 03454. DB-1 Precision Preamp, $397, with DB-2 Power Supply, 878. Five-year warranty; manufacturer pays return freight. Tested #1271128/2271128, on loan from manufacturer.

Since the D B still keeps cropping up in some SOTA discussions, we figured we might as well sock it into our current reference set up and give it another listen.

Well, it still sounds a little hard and zippy. It doesn't actually have a nasty edge; it's not really unpleasant or irritating; it just isn't in the super category. We'd rate it maybe one small notch below the Van Alstine and on a par with, or very slightly ahead of, the AGI 511A.

One very interesting thing about the DB-1 is the way it interfaces with the D B Systems DB-6 power amp. The latter puts a very peculiar step into the leading edge of a square wave passing through it, sort of like a smooth bite taken out of the leading corner. When the square wave is first passed through the high level stage of the DB-1, it's slowed down to the point where it no longer excites this behavior in the DB-6. Other preamps we've tried have no comparable "corrective" effect. We suspect there are some elusive feedback-related transient problems going on in D B equipment that may account for what we hear.

Incidentally, driving the DB-6 with the DB-1 didn't endow the former with more transparency, definition and spaciousness. So we're back to square one with D B-we find their stuff good but not great and we still don't know why.

GAS 'Thalia'

The Great American Sound Co., Inc., 20940 Lassen Street, Chatsworth, CA 91311. Model A-801 'Thalia' Servo-Loop Preamplifier, $299. Five-year warranty; customer pays all freight. Tested #A4-801439, on loan from manufacturer.

We don't like this one at all, although its behavior on the laboratory bench is exemplary.

Close to perfect RIAA equalization, beautiful reproduction of RIAA-preemphasized square waves (very fast rise time!), harmonic distortion of the order of 0.01% at 2 volts out throughout the audio range (even at 20 kHz!), you name it-everything is hunky-dory.

Except the sound. The high frequencies are badly smeared. The midrange is thick and poorly defined. And there's a pervasive haze or veil over the whole sonic presentation, making you want to wipe it off or tear it away.

We're not saying that any of this is irritating or fatiguing. Just inaccurate. The design, we're told, is strictly P.B. (Post-Bongiorno), so we can't even talk about the taste for zabaglione slathered over the sound . . .

If nothing else, this proves the time has come for our new measurement procedures. It's frustrating to see such a chasm between the lab and the listening room.

Hafler DH-101

The David Hafler Company, 5817 Roosevelt Avenue, Pennsauken, NJ 08109. Model DH-101 Stereo Preamplifier, 8299.95 wired. (In kit form, $199.95.) One-year warranty; manufacturer pays return freight. Tested #1805801, on loan from manufacturer.

Here it is, audio fans-the return of The Wizard of Cheap-but-Good. Dave Hafler, the man who conceived the original (tube) Dyna kits, has come out of semiretirement and is bent on showing the world that the best sound money can buy doesn't necessarily mean a whole lot of money, not even in the solid-state era. He en listed the very knowledgeable Ed Gately to de sign the circuit of a totally new and different preamp, complete with tone controls (no stripped down little black box for these classy gents), and they're practically giving it away.

Well, is it SOTA or is it a piece of junk? The answer to that is a resounding "Yes!" When we first listened to the DH-101 through our reference system, we were amazed.

Other than the Mark Levinson and the Hegeman, no other preamp we had ever tried was even in the same league. Completely open, spacious, uncolored, focused and detailed, the sound was a joy and, at $300, an economic miracle. After long and agonizing A-B-ing, we decided that the Hegeman had definitely more authoritative bass and perhaps a shade more definition and immediacy, and that the Mark Levinson was the best of them all-but that was it. Everything else sounded less good than the Hafler, including the excellent Audionics. We then left the DH-101 "cooking" on our equipment rack with the power on (we do this routinely when testing low-wattage equipment) and turned our attention to other things.

Coming back to it a few weeks later for a recheck, we found the DH-101 sounding totally different. Gone was the live, transparent quality; it seemed that some kind of lid had been clamped on the sound. Not that we heard any thing really bad, let alone obviously defective, but the super performance that had excited us was unquestionably gone, leaving the preamp one small notch above the GAS 'Thalia' in overall sonic accuracy. We then turned it off for a few days to let it "rest" and came back to it again. This time it sounded a little better but nowhere near its originally established par, and right in the middle of playing it went blah again, actually sounding a little hard (a la DB) for the first time. That did it.

Obviously what's happening in the Hafler preamp is that certain critical components change their correct value or response as they are burned in. In other words, El Cheapo strikes again. There are a number of electrolytic capacitors in the signal path; they may well be the main culprits. We aren't sure. We do like the circuit itself; it's quite original in concept and makes a lot of sense. But maybe it can't be put together as inexpensively as Dave and Ed originally figured. Too damn bad.

(We'll try to take a look at a second sample to see if the same thing happens. One other user we know had a similar experience with the DH 101.) Our bench tests, which took place before the decline and fall, revealed nothing we could even mildly complain about, except possibly the RIAA equalization. The Hafler is one of the two preamps covered here that follow that "new" characteristic (the other is the Audionics); no defeat switch is provided for the built in low-frequency roll-off, but the bass control can be used in a pinch as a substitute. The trouble is that the topology seems to fall into the aforementioned hump-and-saddle trap, at least on the low end, with a very broad dip of 0.5 to 0.6 dB centering on 150 Hz-a bad place.

On the other hand, the equalization is almost perfect from 1 kHz on up. And everything else --THD, square waves, all the routine stuff-is absolutely shipshape. One of the DH-101's claims to unique excellence is positive/negative pulse symmetry; the claim is true but the test is passed by many other units.

What is unique about the Hafler, at least among the preamps we've tested, is that it can give you an electric shock. Nothing lethal; just a nice, juicy tingle. Two large bypass capacitors on the AC line (0.01 microfarad each) are responsible for the leakage current that does this.

Their purpose is to filter out RFI, but they do make the chassis rather exciting to touch if you happen to be earthed at the same time.

Of course, everybody knows that it's difficult to shock people nowadays; perhaps even a preamp that sounds different on Tuesday than on Monday won't do it, but a milliamp or two of leakage current definitely will.

Hegeman HPR /CU

Hegeman Audio Products Inc. (Hapi), 176 Linden Avenue, Glen Ridge, NJ 07028. Model HPR preamplifier with Model HCU control unit (incorporating power supply for HPR), 3645 complete. Two-year warranty. Tested #103/101, on loan from manufacturer.

"Im a lazy engineer," confesses Stew Hegeman, who at 65 has three times as many original audio designs to his name as any two hyperactive 32.5-year-old engineers in the business. His laziness, he explains, caused him to design just one low level amplification stage, to be used repetitively in all of his current products. It's built around the uA739 op amp chip, compensated just so, and features tremendous bandwidth and high linearity, with out the need for excessive feedback. (The uA739 is a true class A amplifier and probably the only genuine analog chip around.) Stew takes two of these circuit modules, puts a passive RIAA equalization network be tween them, and that's one channel of his HPR phono strip, which goes right next to or under your turntable. The HCU control unit, built around one more of these modules per channel, then goes anywhere you have room for it, as long as the HPR's umbilical cord reaches the HCU?'s built-in power supply. It's really a very simple arrangement, and there's good logic be hind it. The HPR has near-zero input capacitance, and minimizing the cable length between the arm and the input helps maintain that low figure. The low output impedance of the HPR then makes the cable capacitance between it and the control unit irrelevant. We don't think this chassis division is the secret of the Hegeman preamp's excellence, but it happens to be a very nice idea.

The HPR/CU combination turned out to be the most accurate-sounding preamp we had ever tested with the exception of the very latest version of the Mark Levinson ML-1. The Hegeman reproduces a sound stage of realistic depth and uncanny immediacy at the same time, with a genuine see-through quality in the mid range and highly focused inner detail. It's just a clearer and better defined sound than is avail able through other preamps. The highs are a little bit intense (not hard or irritating, though

-far from it), possibly because Hegeman deliberately inserts a ""corner' at approximately 21 kHz to stop the high-frequency slope of the RIAA curve, whereas in the cutter-head preemphasis network the corresponding corner is more likely to be at 40 kHz or thereabouts.

(See also our earlier discussion of this above.) The bass is magnificent; Hegeman believes it's because he carries the equalization down to practically DC. (This is a whole can of worms that we won't open here now.) Another interesting thing is that between about 80 Hz and 6 kHz, the RIAA equalization error is actually zero-a very rare phenomenon, especially in the midrange. Does that have anything to do with the freedom from colorations? We refuse to speculate.

The distortion figures of the Hegeman are far from spectacular. High single-oh and point one-something are typical on most static measurements. Obviously it isn't a feedback-happy design and just as obviously it has low dynamic distortion. Amen.

We must add that Hegeman Audio Products is a very small and only recently formed electronics company, with limited production facilities. We witnessed the slow and painful debugging of preproduction units such as ours, and we haven't seen an absolutely final, this is-it production model yet. In other words, our endorsement of this exciting new product is necessarily conditional.

But if someone asked us right now which is the best preamp in the world at a three figure price, we couldn't give any other answer but the Hegeman.

Linn Moving Coil Preamp

Audiophile Systems, 5750 Rymark Court, Indianapolis, IN 46250. Linn/ Naim Type PNAG moving coil preamp with Type NAPS power supply, $250. Tested #0023/ 240V, on loan from distributor.

This Scottish-made pre-preamplifier for MC cartridges has been hailed in some quarters as a Verion beater, an assertion that (a) implies a somewhat casual attitude toward the laws of physics and (b) is untrue. (See our follow-up review of the Verion transformer below.) The Linn happens to be a very nice-sounding low-level amplification module, preferable in our opinion to the Mark Levinson JC-1 and the equivalent System D for the ML-1. We believe it introduces less veiling and coloration.

On the other hand, the Verion transformer introduces no character of its own whatsoever and presents a much more natural and balanced sound than the Linn. On transient peaks, the Linn is a little hard and glassy by comparison, and in noise level there's simply no comparison.

Because no amplifier is as good as no amplifier at all.

Mark Levinson ML-1 (follow-up)

Mark Levinson Audio Systems, 55 Circular Avenue, Hamden, CT 06514. ML-1 Preamplifier, with plug-in System A3X, $1250. Five-year warranty, customer pays all freight. Tested #2221 (with factory updates), owned by The Audio Critic.

We believe we've found the root of the ML-1 controversy. Some audio purists assert that the ML-1 is close to perfection-and they're right. Others find the ML-1 either hard or closed-down in sound-and they're right.

How come? It's that damn volume control.

To guarantee accurate tracking within a small fraction of a dB, and to exclude the possibility of various volume-control-induced noises and distortions, Mark Levinson uses a ridiculously expensive wirewound potentiometer made by Spectrol. The original design of the pot was for servo control, not audio. A number of these super pots turned out to be defective in various exotic ways. Some of them started out with barely perceptible flaws, if any, and then gradually went sour. We had one of these.

We couldn't quite understand why we had started to have less and less respect for our ML-1.

Maybe we're becoming more critical, we thought, and our reference system is getting better. Well, it was the Spectrol. By the time we finally took a close look, it was making square waves ring with a leading-edge spike that went right through the ceiling. And the highs were biting like sharks.

In furnishing us with a replacement pot, Mark Levinson skipped a step. We were given the very latest, corrected-to-the-nth-degree Spectrol that will go into production units only beginning in April. The last version before that, we're told, was also quite trouble-free though not absolutely guaranteed to be so.

The early model (that was ours) is highly suspect. You can recognize it by its completely smooth housing, without any external resistors soldered on it. If you have one of these, get in touch with the company and they'll take it from there. In fact, if you have any reservation at all about the volume control (or any other part) of your ML-1, don't hesitate to speak up.

The whole point of insanely expensive equipment like this is that the maker can't afford to let you down. The Mark Levinson people are ready to lose time and money to update your preamp if it doesn't meet specs. When in doubt, ask them. They've got all the serial numbers of the bad guys and they want them back.

At the same time as our new volume control was installed, a few other updates were effected. These included new line driver modules (the latest are the best), the new A3X export phono cards (set for a gain of 33 dB to go with our Verion transformer), a couple of new resistors, plus some changes in component location. All of these modifications are standard on current production units except the A3X cards; the domestic A3 cards differ by one resistor and a few dB more gain. The important thing is to have the latest System A, whether domestic or export, as the earlier ones had an RIAA error of +0.4 dB or so humping up in the 10 to 12 kHz region, whereas the new ones are extremely flat (we measured +0.15 dB from 35 Hz to 35 kHz and -0.7 dB at 20 Hz) and continue their high-frequency slope at least three octaves higher than the old ones. (Forget about System D. The Verion transformer blows it away.) With these changes our ML-1 reached new and unsuspected sonic heights. We had never hoped to find a better preamp than the Hegeman, but this one is better. The ML-1 sounds somehow more controlled, tighter and cleaner, with even greater immediacy and sharper focus. By comparison (and only by comparison) the Hegeman "blooms" a little. What we said about the ML-2 power amplifier is equally applicable here: next to the ML-1 all preamps have "hair" on their sound. The definition obtainable with the ML-1 is almost scary; things like double-tonguing in trumpet passages can be heard with an etched clarity beyond all expectation. Of course, some people may prefer that trace of bloom in the Hegeman (we suspect it has to do with a bit of innocuous second harmonic distortion, of which the ML-1 has a whole order of magnitude less-and without feedback instabilities); if so, they're entitled to their opinion and they've just saved them selves $605.

That's the permanent trap Mark Levinson has chosen to build for himself: unless he de livers the best there is, period, his prices make no sense at all. As far as the latest production version of the ML-1 is concerned, we believe Mark has escaped once again.

Rappaport PRE-1A (follow-up)

A.S. Rappaport Co., Inc., 530 Main Street, Armonk, NY 10504. Model PRE-1A Stereo Preamplifier (with PS-1 external power supply), 8755. (Also available with internal power supply as Model PRE-1, $620; with internal power supply and without tone controls as Model PRE-2, $520.) Three-year warranty; manufacturer pays all freight. Tested #1481-01/#2007, owned by The Audio Critic.

Having been apprised of certain internal refinements in the Rappaport, we traded in our early (almost prehistoric) sample and inserted the latest production version into our comparison lash-up. It did extremely well, giving all comers a hard time in A-B tests, but this time it didn't end up among our final recommendations.

The only flaw of the PRE-1A is a barely perceptible thickness-or call it nasality-in the midrange and the uppermost region of the bass. It's difficult to zero in on but it's there all right; you can hear it on brasses, for ex ample. The result is a slight recession of realism and immediacy, a kind of homogenization, at least in comparison with the Mark Levinson and the Hegeman. The Audionics, too, has just a bit more midrange openness and transparency, but the Rappaport beats it on the top end. In fact, the Rappaport has become ultra smooth and listenable in the upper octaves; we prefer it even to the Hegeman in that respect (though not to the Mark Levinson). All in all, a marvelous preamp still, barely squeezed out by the new generation.

It must be pointed out that, among the super preamps, the Rappaport has the least ac curate RIAA equalization curve; the error is +0.7 dB in the 5 to 10 kHz octave and the high frequency slope is flattened out only an octave or so above the audio range. The funny thing is that this ought to result in a subjective characteristic tipped toward the top end-but, if anything, the sound of the PRE-1A can only be called bassy, although the 20 Hz response is -0.7 dB and the midbass is dead flat. So the little anomalies we heard have nothing to do with amplitude response (Mark Davis, please note).

On static distortion measurements, the PRE-1A appears to be impeccable.

Currently we also have a sample of the new PRE-2 (same thing but with internal power sup ply and no tone controls), but it arrived too late for these tests. Our superficial impression of it is that it's ever-so-slightly better than the PRE-1A, possibly owing to the simpler, more directly wired signal path. We plan to report on it in the next issue.

One thing you can be sure of. This isn't Andy Rappaport's last preamplifier. At the ripe old age of 20, he speaks of the PRE-1 fondly as his good old preamp and is working on something so far out he doesn't even want us to talk about it.

Van Alstine Model One

Van Alstine Audio Systems, Inc., 2202 River Hills Drive, Burnsville, MN 55337. Model One direct-coupled stereo preamplifier, 3600. Three-year warranty. Tested #173101, on loan from manufacturer.

This is a very decent preamplifier, al though it doesn't quite live up to the mystique with which its makers try to imbue it. We have a private communication from Frank Van Alstine that speaks of "our design-no feedback, no TIM, no phase shift" but the circuit details are kept very hush-hush and the epoxy-en capsulated plug-in modules withhold all in formation except name, rank and serial number. We seem to recall some mutterings about hybrid IC's, especially made for the Model One. We wonder.

The sound of the Van Alstine is beautifully open and essentially neutral, but there's a certain degree of grainy hardness and zippiness that keeps it out of the winner's circle. At first we thought this was due to a defect, as we had noticed too much DC offset in one channel; that was immediately fixed by the manufacturer and the whole unit rechecked.

Well, it still sounds just a tiny bit edgy, grainy and irritating. Not unlike the D B, although the Van Alstine sounds more solid and authoritative on the whole. We'd rank it somewhere in the middle of the group reviewed here and unequivocally behind the Mark Levinson, Hegeman, Audionics and Rappaport.

In the laboratory we admired the almost incredibly accurate RIAA equalization, found everything all right on static distortion measurements, and were baffled by a weird trailing edge spike that the phono stage puts on RIAA preemphasized square waves. Since circuit de tails aren't available, we can only speculate about the relationship between this transient handling peculiarity and the slight edginess of the sound.

None of this should be taken as a put down of the Van Alstine; if we had nothing else to compare it with, we could live with it contentedly. The nature of our business is such, however, that we do have something else to compare it with.

Verion Mark I (follow-up)

Verion Audio Inc., 75 Haven Avenue, Mount Vernon, NY 10553. Stereo Pickup Transformer Mark I, $375.

Five-year warranty. Tested #1S224598 and #15329497, owned by The Audio Critic.

After a year of shuffling moving-coil cartridges in and out of our reference system, and experimenting with devices to match them to our phono input, we want to make it loud and clear where we stand on this subject in general and on the Verion transformer in particular.

"Of course, any transformer in the signal path is a no-no," writes one of our correspondents, thus epitomizing the untutured folklore of audio-store cowboys. Little does he suspect that a transformer is actually the theoretically ideal impedance-matching device and that all you need in the case of an MC cartridge is impedance matching, as the energy coming out of the cartridge terminals (voltage times cur rent) is certainly enough to drive any phono input circuit. The fact that the audio industry has traditionally produced crummy transformers has nothing to do with this fact of physics.

It so happens that making a textbook-perfect transformer is extremely costly.

First of all you have to make sure that the core can't possibly saturate at the highest signal levels anticipated. That usually means that the transformer must be made bigger than any one figured in his wildest dreams. Then you must deal with nasty problems like leakage inductance, shunt capacitance and hum shielding.

All that involves a fairly sophisticated technology but no defiance of the laws of nature.

Once you have placed all reactive interactions with both source and load well outside the de sired bandwidth of the transformer, and kept externally induced hum and noise out of the signal path, you have a perfect passive device that must let more information through within its bandpass than any electronic circuit, for the same reason that empty eyeglass frames must let more light through than eyeglasses. There's no point arguing about this around the water cooler; it isn't negotiable.

What Mitchell Cotter did when he de signed the Verion Mark I was simply that he went all the way and ended up with a complete and sufficient transformer (for phono signal levels and a bandwidth of approximately 80 kHz) rather than an under-designed and in complete one as others did historically before him. As a result, the Verion simply wipes out pre-preamps and head amps with respect to noise level and handily outperforms them with respect to signal fidelity-not because Mitch 1s a magician but because the laws of nature are on his side.

We know of absolutely nothing the audio purist can put between his MC cartridge and his preamp input that will let the signal through with as little audible alteration, either additive or subtractive, as the Verion transformer. Not the Mark Levinson JC-1 or System D, not the Rappaport MC-1, not the Linn, not the DB-4, not the MAS 1, not the Dayton Wright DW 535, not the GAS 'Goliath'-nothing. Make sure, though, that you get the correct strapping for your particular MC cartridge (P, S, PP, or X, as specified in the Verion literature), other wise you'll lose a couple of dB of signal-to noise ratio and dynamic range.

And if you're buying an expensive preamp (whether reviewed by us or not), don't buy it with the moving-coil option if you have a choice. The extra circuit won't be as good as the Verion. Mother Nature says so.

Recommendations

When it comes to preamplifiers, our usual disclaimer about not having tested them all be gins to sound a little hollow. We've tested an awful lot of them, and it would take something quite new and special to replace the choices be low. Not that it can't happen; we just don't expect it to.

Best preamplifier so far, regardless of price: Mark Levinson ML-1 (latest updated version only-see review).

Close to the best at a much lower price (and, incidentally, the next best regardless of price): Hegeman HPR/CU.

Best preamplifier per dollar: Audionics BT-2 (for absolute ranking, see review).

Best way to play moving-coil cartridges: Verion Mark I.

---------

[adapted from TAC]

---------

Also see:

The Great Preamp Survey

The Great Preamp Survey: Part II

Box 392: Letters to the Editor

Various audio and high-fidelity magazines