The Ever-Changing Preamplifier Scene (Vol. 2, No. 3, Spring - Fall 1980)

The Ever-Changing Preamplifier Scene: Audio Research SP-6B (interim report) , Beveridge RM-1a/2a , Cotter System 2 (reappraisal) , Mark Levinson ML-7 (interim report), Precision Fidelity CTA , Robert Grodinsky Research Model Four

By the Staff of The Audio Critic

Now what? A $650 full-control preamplifier that sounds as good as, or possibly even a little better than, any other at any price. That's what. Plus a number of other goodies.

Nothing has happened since the last issue to change our views from the ones stated there in the introduction to the preamp reviews (Vol. 2, No. 2, p. 53). We still believe that bench tests are of very limited value in separating good, better and best, even though they are very useful for identifying obvious design errors and isolating actual defects. We also continue to perceive a steady convergence toward the same kind of sound, which is ostensibly the uncontaminated sound of the program material itself, by the very best preamps.

(Not all preamps, nor even all preamps having correct amplitude response, as wrong headedly trumpeted by certain populist pundits.) A word about bypass tests.

Since fanatically careful listening tests, performed by insertion into a thoroughly familiar chain of components of the highest possible resolution, are the only way to ascertain audible differences between good preamps, it may be helpful to point out once again the pitfalls of one particular type of listening test, which is especially favored by its advocates for evaluating preamps. This is the straight-wire bypass (some times called SWB by those who regard alphabet soup as a gourmet dish). The sound of the device under test (the ' 'black box') is compared with that of a straight wire that bypasses it, leaving it out of the chain. In other words, it's an A-in/A-out test, rather than the usual A/B test. The line amplifier of a preamp is especially easy to bypass, since it incorporates no equalization, and what little gain it provides can be easily obtained elsewhere in the chain, if at all necessary. But even a phono stage can be pre-equalized to be flat, fed from another signal source, and then switched in or out. In fact, any purely electronic signal path can be bypassed as long as the gain of the total chain is controllable in some way.

So far so good; we've done it ourselves many times and it does provide useful information. The great fallacy of the faithful straight-wire bypass tester, however, is the claim of infallible objectivity. He brags, "Look, Ma. no value judgments! I don't need to decide which sounds better. the straight wire or the device under test: if they sound different I know the device is inaccurate, since the straight-wire sound is by definition the correct sound, right?" Wrong. And, what's worse, scientifically naive.

Let's take an extreme case in order to make our point swiftly. Suppose the black box being bypassed incorporates all the RFI filtering for the signal path. And suppose the straight-wire sound is consequently full of crackles. hash, CB interruptions (''Breaker two, breaker two!'") and as sorted other garbage. With the black box switched in, presto, there's only music. Since they sound different. obviously the straight wire is right and the black box is wrong . . . hey. wait a minute. . . Okay, we know that's an absurd example, but it states the basic philosophical issue. Certain stages of the signal path aren't supposed to be straight-wire-like.

The typical SWB blunder is a bit less obvious. Let's assume that the tester's reference system incorporates a DC-to-light type of power amplifier without any bandwidth limiting filter at its input. (Guys like that are inclined to choose amplifiers like that.) And let's further assume that the device under test is the line amplifier of a preamp (from "'aux'' to 'main output'') and that the rise time of this line amplifier is a fairly slow 3 microseconds. With the line amplifier switched in, the power amp is well protected against out-of-band spikes of energy that might cause the TIM/ SID/DIM type of distortion so many ultrawideband amplifiers are prone to. Very little above 120 kHz or so can get in.

With the straight-wire bypass switched in, the power amp is at the mercy of the program source. The door is open to out-of-band excitations over the unlimited passband of the straight wire. If the power amplifier can't take these, the result is that zingy quality on top so often mistaken for "'detail"' by those who were raised on amplified sound instead of live concerts. The line amplifier may quite possibly pass audio information with total accuracy, but of course it will make the power amp sound smoother than the straight wire connection did and therefore create an audible *'difference." Our tester will then, with full confidence in his objectivity, conclude that the line amplifier is at least slightly inaccurate or ''veiled."" Sancta simplicitas! The moral of all this is not that straight-wire bypass tests are useless. They happen to be highly revealing if you know how to interpret them. But there's no way on earth to eliminate value judgments in listening tests. Regardless of your methodology, you can't escape from judging subjectively which one of two sounds appears to sound more like music.

Or at least more like what you believe to be the true sound of the input. And such a belief can be formed only by listening first to the output of a familiar reference system driven by that input.

Which is where we came in.

Audio Research SP-6B (interim report)

Audio Research Corp., 6801 Shingle Creek Parkway, Minneapolis, MN 55430. Model SP-6B vacuum-tube preamplifier, $1495. Tested samples on loan from owners.

The "'B'' modification represents several updates since the original SP-6 tube preamp we reviewed in Vol. 1, No. 6, but close to press time comes word that this version, too, is about to be succeeded by a further improved one; whether it's a lightly revised B or sufficiently new to be a C, we don't know. This is, of course, the Audio Research way of doing business, but we aren't going to get too deeply involved in the whole thing, anyway, since our exposure to two different SP-6B samples was relatively brief and inconclusive. When we can get our hands on the '"'final'' version, assuming it stays put long enough, we shall analyze it at greater length.

Of the two samples we looked at, one (which happened to be right out of the sealed factory carton) didn't sound very good; it had a hard, pinched, irritating quality that took a minute or two to become aware of but was unmistakably there. The other sample sounded gorgeous, possibly superior to the Cotter PSC-2/CU-2 we were trying to A/B it with, but trying was as far as we got. Both samples, good and bad, proved to be impossible to audition critically for any meaningful length of time because their automatic muting was being falsely triggered by the tiniest warps and bumps in the record (instead of catastrophic pulses only, as intended), so that the music dropped out and came right back again several times per minute, creating a kind of hiccuping effect. This automatic muting feature cannot be regulated nor defeated externally; furthermore, a third and a fourth sample were also reported to us by their owners as suffering from the same defect.

Conclusion: the SP-6B, at this particular stage of its evolution, is somewhat unpredictable in (1) sonic performance and (2) operating reliability. There exists enough vari ability in vacuum tubes, even the premium kind, to account at least partially for the first uncertainty; the second is inexcusable. (Remember-one of our original SP-6 samples in 1978 was also defective.) We also have some unalloyed good news to report, however: the RIAA equalization is vastly improved in the B version of the SP-6; in fact it's very close to being right on the button. Little by little, using the consumer and the equipment reviewer as its free R & D and burn-in facility, Audio Research may be getting its act together. For all we know, there could already be one or two samples of the SP-6B out there that will equal or surpass any other preamp. At this point, we find ourselves unable to get terribly excited by that possibility.

Beveridge RM-1a/2a Harold Beveridge, Inc., 505 East Montecito Street, Santa Barbara, CA 93103. RM-1a Preamplifier with RM-2a Power Supply, $3500.

Five-year warranty (excluding tubes). Tested RM-1 #127, RM-2 #071, preproduction RM-1a and RM-2a, all on loan from manufacturer.

This is the all-out, damn-the-expenses tube preamplifier designed by Roger Modjeski (who made sure his name would be etched right on the circuit board) and manufactured by Harold Beveridge. It started out in the RM-1/RM-2 version at $2150, went up to $2500, and now with 'a' suffix is going right through the roof at $3500. We find that last figure totally alienated from reality, regardless of all the costly features, which include a complement of 12 noise-selected E88CC/6DJ8 tubes in an extremely complicated circuit on an incredible single board, gobs of military-grade parts, and above all a separate solid-state power supply, as big as a full-fledged power amplifier, for individually regulating 6 stages in the preamp. Even so, the mechanical engineering of the chassis left us unimpressed, with sloppy screw holes, misaligned metalwork, cheap hardware, indifferent finish and optional plastic glue-on feet supplied on a strip (no, we aren't kidding); furthermore, the volume control is several grades below the heavy-duty broadcast-type step attenuator that belongs in equipment at this price. We could go on and on, and the manufacturer could on the other hand also rationalize each little imperfect detail; the point is, though, that no manufacturer ever had a gun held to his head to make a$3500 preamp in the first place, so if he does make one we expect it to be of ineffable perfection both sonically and physically.

After all, it's still only a phono amplifier plus a line amplifier with control facilities.

Well, what about that sonic perfection? To take it chronologically, we never had a chance to test the original $2150 version, which was made with Mylar capacitors. We were sent the next version (still without the 'a' suffix but now $2500), in which polypropylene capacitors were used instead of Mylar. This version, of which there may still be a few left on dealers' shelves, sounded absolutely dreadful- hard, strident, with peculiarly ''angry'' highs and a spatially confused midrange. Bench tests offered no clue to these distortions, except for a very small and probably irrelevant overshoot on square waves through the line amplifier under certain conditions. Polypropylene capacitors, unless disastrously defective, couldn't have been the cause either, and yet we kept hearing from other users in the field whose ears we respect that this new version was a big comedown from the original. The mystery remains unsolved to this day; just remember that in the polypropylene version you see mostly yellow capacitors on the circuit board, whereas in the Mylar version a bright orange predominates. Our efforts to deter mine how and why a golden-ear oriented company released this product for sale failed to produce a satisfactory answer.

Finally the company yielded to the pressure of discriminating listeners and went back to the Mylar version, adding some minor refinements along the way and raising the price to $3500. This latest modification is now called the RM-1a/ RM-2a, of which we tested a preproduction sample. What a difference! Beautifully open sound, tremendous definition and dynamics, smooth and airy highs without a trace of stridency, fatigue-free listening for hours. In comparison with the Cotter PSC-2/CU-2 combination, the Beveridge sounded definitely smoother on top without any sacrifice of detail (but see also the Cotter reappraisal below), more buoyant and alive in overall musical impact, at least equal and possibly superior in transparency, but perhaps less tightly controlled and unshakable under the impact of violent transients. The latter quality may or may not have had something to do with the way we had left the adjustable gain controls for the various stages on the circuit board of the Beveridge; with everything adjusted for minimum gain instead of ' 'normal,' there might have been even more dynamic headroom. Unfortunately, one channel went completely noisy on us before we could conclude this part of the test; since the defect wasn't due to a tube, we had no time to troubleshoot it and repair it before going to press. In any event, a $3500 preamp should be immune to that kind of failure, and besides we ended up preferring the sound of the incomparably less expensive Robert Grodinsky preamp (see review below) to that of either the Beveridge or the Cotter.

On the lab bench, the RM-1a/RM-2a delivers all the promises of the spec sheet, including highly accurate RIAA equalization. Roger Modjeski is a perfectionist when it comes to measurable performance characteristics, and his design shows it. We aren't satisfied, however, with (1) tube preamps as a predictable and permanently viable species in general (see also the Audio Research SP-6B review above), (2) the professionalism evidenced by Harold Beveridge, Inc., in the course of this product's brief but checkered history, and (3) the marketing philosophy that no price is too high if some body out there can be enticed to pay it. As a laboratory exercise by an adventurous designer, the Beveridge preamp fascinates us intellectually and aurally. As a product for sale to the consumer, it leaves us with serious doubts.

Cotter System 2 (reappraisal)

Mitchell A. Cotter Co., Inc., 35 Beechwood Avenue, Mount Vernon, NY 10553. System 2, complete ''front end'' consisting of: MK-2 Moving Coil Pickup Transformer, $550 or 3650 (depending on type); PSC-2 Phono Signal Conditioner, $550; CU-2 Control Unit, $1750 (projected price of future production version); NFB-2 Noise Filter/Buffer, $500; PW-2 Master Power Supply, $450. Five-year warranty. Tested samples owned by The Audio Critic, including limited-production ''engineering model' of the CU-2.

The five modules of this complete ''front end'' are now priced at a total of $3800 or $3900 (depending on your choice of transformer type); out of that the PSC-2/CU-2/PW-2 combination, which is functionally equivalent to the other preamplifiers reviewed here, comes to $2750 instead of the originally projected $1800; it would seem, therefore, that a reappraisal is in order, especially since the Cotter units have been our unequivocal reference standard for more than a year and a half. After considerable soul-searching, we have concluded that the Cotter System 2, as a total entity, can no longer be recommended to the audio purist as the ultimate front end for a stereo system and should therefore be "'delisted.'" Here are the particulars of this conclusion:

1. Our opinion of the Cotter transformers remains as high as ever (see also the section on MC step-up devices in this issue). In fact, the Cotter-designed Verion MK-1 trans former, still advertised in the back pages of Audio at the special liquidation price of $350, looks like a major bargain to us-if you can get a properly quality-controlled unit.

2. The CU-2 control center, which is what makes the System 2 a system, still exists only in the form of the ugly and rather sloppily assembled * 'engineering model," of which only a very few were made. After a year and a half, the production model is nowhere in sight; as we go to press we possess absolutely no evidence that even a mock-up of it exists, let alone finished units.

3. The other components of the Cotter system, al though officially in production, are being produced so slowly, in such minute quantities, that consumers may find it difficult if not impossible to obtain them. This company has been beset by more than ordinary problems, and we have no way of predicting how future demand will be met.

4. The sound of the PSC-2/CU-2/PW-2 combination is no longer in a class by itself. Several preamplifiers we recently tested gave comparable or even somewhat superior results. (See the Audio Research SP-6B and the Beveridge RM-1a/2a above, the Mark Levinson ML-7 and the Robert Grodinsky Research Model Four below.)

The possibility exists that an extremely gradual deterioration of one or more of the Cotter units has taken place, in such tiny daily steps that we were at no time aware of it. Looking back, we do recall what seemed like a smoother, sweeter, more perfectly edgeless sound at first; however, memory is a notoriously unreliable witness and the preamp today still sounds and measures outstandingly fine, so we can't really tell whether progress or decay got the better of it. The latter possibility may be more palatable to the theorist who conceived the circuit; to the consumer the end result is exactly the same either way. A $2750 preamp isn't supposed to change.

5. The NFB-2 filter/buffer is gradually becoming obsolete as more and more power amplifier designers realize that their input circuits must incorporate some kind of judicious ultrasonic filtering. Subsonic filtering is also coming into much wider use, especially as a by-passable preamp control but in many cases also at the input of power amps.

Needless to say, two bandpass filters in tandem would be excessive. It could be argued that the filter networks in the NFB-2 are more sophisticated than what others use, especially the elegant high-order filter in the low-pass section, but in practice we have found that a correctly deployed single pole filter as used by the new generation of TIM/SID conscious power amp designers will do the job. Certainly, the era in which 9 out of 10 power amplifiers sounded better with the NFB-2 is rapidly coming to an end. As a matter of fact, our latest listening tests with program sources that preserve the utmost detail (direct-to-disc records, line-contact styli, etc.) and through high-resolution speakers (ribbons, electrostatics, etc.) indicate that the NFB-2 will in some cases remove that very last degree of definition and spatial information. In other words, a little bit of the baby is thrown out with the bath. As long as the bath water was really filthy that was okay, but not anymore.

6. The physical inconvenience of a five-module front end, with its unavoidable tangle of cables and unusual traffic problems on the equipment shelf, can be justified only by overwhelming sonic superiority, which appears to be no longer the case.

Mark Levinson ML-7 (interim report) Mark Levinson Audio Systems, Ltd., 131 Leeder Hill Drive, Building 261, Hamden, CT 06517 (or PO Box 6183, Hamden, CT. 06517). ML-7 Preamplifier, $4000 to $5000 (depending on phono options chosen). Five-year warranty. Tested sample on loan from dealer.

We had a chance to spend just a few hours in our laboratory and listening room with 'Mark Levinson's definitive statement on preamplifiers'" (Bert Whyte's words) and can only report the following:

The physical construction of the preamp, which looks like a slightly deeper ML-1, appears to be of jewel-like perfection; most high-end preamps look slightly unappetizing next to it. That has nothing to do with the quality of circuit components, where several other units may be on a par, but with the loving care shown in the planning of all external and internal details.

We liked the sound of the L2 (low-gain) phono module better than that of the L3 (high-gain), which eliminates the need for MC step-up devices. With the L2's inserted, the ML-7 sounded smoother, more at ease and even less fatiguing than the Cotter PSC-2/CU-2, although the difference wasn't devastating; we would have needed more time to evaluate which had the better focus and detail. That was the only A/B comparison we were able to make.

On the lab bench, we had time to verify only that the RIAA equalization was accurate; this had been a point of laxness at MLAS in the past. All in all, a SOTA contender--and at $4000 it had better be. We're hoping to obtain another sample for more thorough testing in the near future.

Now that the ML-1 is being gradually phased out to be replaced by the ML-7, it may be almost irrelevant to report that we came across a sample of the ML-1 that was much better than the one we had owned and reviewed. This was also a fairly old unit, with the older PLS-150 power supply, but it had been completely refurbished at the factory in mid-1980, with all the latest tweaks. This, too, sounded somewhat smoother and more ''musical'' than the Cotter, without any trade-off in inner detail; how it would have fared against the ML-7 we have no idea. Thought you'd like to know.

Precision Fidelity C7A Precision Fidelity, 1238 Green Street, San Francisco, CA 94109.

C7A cascode preamplifier, $549.95. One-year warranty (tubes 90 days). Tested #80501, on loan from manufacturer.

This is a very lightly revised version of the C7 tube preamp reviewed in the last issue; the differences are trivial.

The sound still has that pleasantly thick, dark, cushioned quality, with excellent dynamics, that we found in the C7; if anything, the C7A exhibits the same traits to an even greater extent, with perhaps slightly more control and solidity. True transparency it ain't, however, as an A/B switch to the Robert Grodinsky preamp ($100 more) will prove in about three seconds. (See review below.) As we said before, if it weren't for the top solid-state preamps, we could grow very fond of this neat little vacuum-tube unit.

The RIAA equalization is still off by -0.5 dB in the lower midrange (dip centering on 175 Hz); in one channel there has also appeared a new and unexpected boost of 0.7 dB at 20 Hz. From 1 kHz on up, everything is still perfect. We also find the tolerances in the ' 'low filter'' response a little on the sloppy side. Overall, though, the C7A illustrates perfectly the old French saying: 'The more it changes, the more it's the same thing."' Robert Grodinsky Research Model Four Robert Grodinsky Research, a division of RG Dynamics, Inc., 4448 West Howard Street, Skokie, IL 60076.

Model Four Stereo Control Preamplifier, $650.

Two-year warranty. Tested preproduction sample, on loan from manufacturer.

Once in a blue moon, routine testing can turn into a major event for the equipment reviewer. This is one such case, but first a word of warning:

Physically, the Model Four looks almost identical to the older RG Dimension 3 preamp, which is by now a known quantity in the marketplace. (Audio, for example, reviewed it in its May 1980 issue.) Electrically, however, the Model Four is more sophisticated. Among other things, a cascode circuit has replaced the previous configuration at the input of each amplifier stage. The power supply has also been beefed up, right up to the point where a larger chassis would have been needed to go any further. Thus the two preamps are not the same. Although the now discontinued RG D3 was an excellent unit (we tested it but see no reason to review it anymore), the Model Four is considerably better, so much so that RG Dynamics has decided to market it under the new Robert Grodinsky Research name.

Now for the mindblower. In direct A/B listening comparisons with the best preamplifiers, solid-state or tube, regardless of price, the Model Four came out on top every time, at least by a small margin. Some of our listening panelists went only so far as to find it *'just as good'' as any other; the majority said ''better''-and at $650 that's coming out on top in our book. The principal reason for sonic superiority is nearly always superior circuit design, and good thinking costs no more than bad thinking (one of our dogmatic tenets), so the whole thing isn't necessarily a miracle.

Good parts do cost more than run-of-the-mill parts, however, and the Robert Grodinsky preamp is no Mark Levinson in that respect. It's very decently made, with the money spent in the right places, but we wouldn't put it in the Voyager spacecraft. On the other hand, this is unlikely to be still our reference preamplifier (or yours) in 1990, so why worry now? The fact remains that it sounds as transparent and spacious in the midrange, as rock-solid on the bottom, and as clean, smooth and detailed on top as any preamplifier known to us and then some, from phono input to main output. Phono noise may be a smidgen higher. in level than would be ideal, but not high enough to be bothersome. The dynamic qualities of the Model Four are especially remarkable; it never seems to protest under the impact of heavy transients but remains pleasant and un-ruftled at all times.

As a matter of fact, the chief performance claim made by Bob Grodinsky for his circuit is the ability to handle extreme peaks in the phono stage without overload. He has even devised a special torture test, which we may also phase into our bench routine, to prove his point. His test signal is an RIAA preemphasized 500 Hz square wave, passed through a single-pole filter (-3 dB at 60 kHz) and ending up with a rise time of 1 microsecond. He feeds this spiky waveform into his phono stage at 3750 millivolts peak-to-peak, to simulate a crest factor and speed well beyond that of real-world pickup outputs. We aren't saying that no other preamp will pass this signal without distress, but his certainly does. We also found conventionally measured square-wave response to be close to perfection through both the phono stage and the line amplifier. As for the RIAA equalization, the error is possibly the lowest we have ever measured, very close to +0.0 dB.

Incidentally, the RIAA deemphasis in the Model Four is passive at the higher frequencies and active at the lower, an interesting solution that has some distinct advantages. So even for techno-one-upmanship you don't absolutely have to buy a $2000 or $3000 preamp.

What's more, your $650 doesn't just buy a stripped down, straight-through, minimal preamp in this case. The Model Four has every conceivable switch, control and facility, including bypassed tone controls, subsonic filter and two buffered tape loops. About the only thing we miss is an absolute phase switch, a la Beveridge or Cotter. That would be very useful in our Reference A system, which is where we have the Robert Grodinsky preamp as we go to press.

Recommendations

We're fully aware of the credibility problem we may be creating here, but we've got to call them as we see them.

Best-sounding preamplifier tested so far, regard less of price: Robert Grodinsky Research Model Four.

Best preamplifier per dollar: Robert Grodinsky Research Model Four.

'Reference A' Update As our regular readers know by now, the ''A'' designation stands for the sonically most revealing and musically most believable stereo system we're able to assemble at any given time, out of the components tested and reviewed up to that time. Price as such is no object, although frantic over spending for nonmusical reasons is emphatically not our style. Insertion into Reference A is always our first step in the comparative listening evaluation of new components that come in for testing. These updates will be a permanent feature of our forthcoming biweekly Bulletins.

Speaker System

From 100 Hz on up, the Quad electrostatic loud speaker ($1780 the pair). Below 100 Hz, the Janis W-1 subwoofer ($1450 the pair). For greater flexibility (though not necessarily better sound at the ''sweet spot' of the Quads), the Pyramid T-1 ribbon tweeter ($1175 the pair) is an add-on option.

As an alternative system at a huge saving, the new Fourier 1 full-range speaker ($1190 the pair) comes surprisingly close to the above in overall performance characteristics.

Power Amps and Crossovers

To drive the Quads, the Bedini Model 25/25 ($845);

to drive the Janis subwoofers, a pair of Janis Interphase 1A bass amplifiers with built-in 100 Hz electronic crossovers ($1130 the pair); to drive the optional Pyramid tweeters, another Bedini Model 25/25 ($845). The Pyramid incorporates a passive high-pass filter; an electronic crossover would be greatly preferable, but we're still experimenting to finalize our choice.

To drive the Fourier 1 full-range speakers, a pair of The Leach Superamp mono units ($1598 the pair).

Preamplifier and MC Step-Up

Robert Grodinsky Research Model Four ($650), with Cotter MK-2L transformer ($650).

Phono Cartridge

Fidelity Research MC-201 ($325).

Turntable and Arm Win Laboratories SDC-10 ($2925) and SDA-10 ($1000).

The base price of Reference A is now down to $10,755, as a result of some elegant simplification without sonic compromises. The ribbon tweeter option with the extra power amp (but without electronic crossover) brings the price up to $12,775. The '""'economy'' version with monolithic speakers costs only $8338.

-----

[adapted from TAC, Vol.2, No.3 ]

---------

Also see:

The Cartridge/Arm/Turntable Scene: Credible Improvements at Incredible Prices -- EON POD Disc Clamp, Fidelity Research MC-201, JMAS MIT-1 ; Platter Matter , Win Laboratories SDC-10, SDA-10, SDT-10-2 and SPG-10

Records & Recording-- A Few More for the Demo Shelf

Various audio and high-fidelity magazines