Box 392: Letters to the Editor (Vol.1, No.6: Spring-Fall 1978)

By now our editorial correspondence has fallen pretty much into a pattern. We receive basically three categories of letters. The first consists of the I-love-you-you're-the-greatest and the get-lost you-creeps varieties. (Not in equal numbers, we must hasten to add.) We publish very few of these, since there usually isn't much that our subscribers can learn from them. The second category addresses the various broad issues in audio on which we've taken sides and includes affirmations, disagreements, additional information and suggestions. These we publish whenever it appears to us that the letter writer knows what he is talking about or else represents an interestingly erroneous point of view that we consider worth refuting. Lastly, there are the letters from manufacturers in response to our equipment reviews. Our policy is to publish these without restriction, unless they are totally scurrilous and devoid of information. The letters we publish may or may not be excerpted, at the discretion of the Editor. Ellipsis (. . .) indicates omission. Address all editorial correspondence to: The Editor, The Audio Critic, Box 392, Bronxville, New York 10708.

The Audio Critic:

In recent issues of The Audio Critic there have been several references to the need for low feedback factor and wide open-loop bandwidth in audio power amplifiers to avoid Transient Inter modulation Distortion (TIM). Although several papers have been written advancing such arguments, the subject is still quite controversial, and several authors have expressed opposing views on the subject. I would like to explain the following points: 1) feedback factor and open-loop bandwidth are, by them selves, irrelevant to TIM avoidance; 2) slew rate is the most important single key to TIM avoidance; 3) it is a common misconception that slew rate is smaller in amplifiers with high feedback factor and narrow open-loop bandwidth.

A typical explanation of TIM goes something like this: Feedback amplifiers operate on the principle that a large portion of the input signal is cancelled by feedback from the amplifier output, leaving a small signal-plus-error which drives the amplifier so as to produce the desired output. In amplifiers with large amounts of negative feedback, this signal-plus-error is forced to be very small and thus, in theory, low distortion results. The large feedback factor is obtained by putting a large amount of gain in the forward path of the "open-loop" amplifier. The open-loop amplifier is thus very sensitive and in fact will over load if the error gets at all appreciable.

All real amplifiers have finite delay.

If a feedback amplifier is driven with a signal having a very fast rise time (like the leading edge of a square wave), there will be a brief period during which the open-loop amplifier sees the full input signal, undiminished by negative feed back which hasn't gotten there yet. Over load will thus occur and distortion will result. The more sensitive inputs of amplifiers with high feedback factors are that much more prone to such overload.

The situation is further aggravated by the slowness of response in the open loop amplifier introduced by necessary feedback compensation. Each stage in a multistage amplifier contributes phase shift, which increases with frequency.

Feedback compensation rolls off the open-loop frequency response so that the feedback factor falls below unity before enough excess phase shift accumulates to cause instability or peaking in the closed loop response. In most amplifiers the gain is rolled off at 6 dB per octave with a single dominant pole. The frequency where the feedback factor falls to unity is called the gain crossover frequency and is usually in the vicinity of 1 MHz for power amplifiers. This constraint is primarily due to the power transistors, which have ft's of a few MHz and begin to contribute substantial phase shift above 1 MHz.

Amplifiers with large feedback factors have more gain to get rid of, and must start their compensation roll-off at a lower frequency, resulting in narrower open-loop bandwidth. An amplifier with 20 dB of feedback needn't start its roll off until 100 kHz for a 1MHz gain cross over, while one with 60 dB of feedback must start at 1 kHz. The latter amplifier, with heavier feedback compensation and a long open-loop time constant, is thus slower in responding to input signals. In responding to a fast input signal, a large internal voltage or current overshoot will be produced in order to charge the compensating capacitor quickly and overcome the effect of the long time constant. In some cases this overshoot may be 1000 percent com pared to '"'nominal" signal levels. If the overshoot causes stages prior to the compensation to clip or become nonlinear, TIM results. When these overshoots are clipped, the amplifier is into slew-rate limiting; TIM is thus sometimes referred to as "slewing induced distortion." It has been shown mathematically that if the in put signal is band-limited to a frequency less than the open-loop bandwidth, no overshoot can occur. Thus, wide open loop bandwidth eliminates the possibility of TIM caused by overshoots.

Having wide open-loop bandwidth, in turn, places a limit on the feedback factor.

The above explanation seems plausible enough and has been published in many forms in many places. It is the origin of the popular belief that small feedback factors and wide open-loop bandwidth are necessary for minimizing TIM. Although some of the papers on TIM have examined the above arguments in great detail and precision, some crucial considerations which alter the conclusions have been left out or dealt with inadequately. These include: 1) rise times of program signals; 2) propagation delay of real amplifiers; 3) design factors affecting overshoot magnitude and slew rate. There also seems to be a tendency to confuse rise time, rate-of-rise and delay.

The question regarding large-signal rise times of real signals is probably the most controversial, but we can get a handle on it by estimating the ratio of peak slew rate to peak amplitude for program signals. Assuming a phono source, we first recognize that for frequencies be tween about 500 and 2000 Hz, post RIAA amplitude and recorded velocity are directly related. Above 2 kHz, post RIAA slew rate and velocity are directly related because of the integrating effect of the 2 kHz RIAA roll-off. Assuming realistic values for maximum midband and high-frequency velocities, we can then obtain an approximation to the ratio. If we are very generous by using a small midband value of 25 cm/sec and a large high-frequency value of 150 cm/sec (probably impossible to track), we arrive at a figure of 0.076 (V/uS)/V. Thus, a 100-watt amplifier must cleanly reproduce slew rates of 3 V/uS. I know, it sounds very small. Remember, I said cleanly. Keep in mind that although something like a cymbal crash has energy out to 200 kHz, its spectral density be gins to roll off at 12 dB/octave at about 15 kHz.

The open-loop amplifier can be molded as a block of gain, a first-order R-C compensation roll-off, and a delay.

After a sudden change in the input, the output will begin to change after the delay. The time constant of the R-C compensation does not affect this time interval. The closed-loop amplifier is a linear system so long as these elements are linear. It is also a continuous system in spite of the delay. Feedback is present 100% of the time as long as no stages are clipped. There are no intervals during which the amplifier operates open-loop and exhibits increased gain. The feed back is, however, continuously "out of date" by a time equal to the delay. Fortunately, the feedback equations readily take this into account in both the frequency and time domains. Based on stability considerations, it can be shown that this delay must be typically less than 125 nanoseconds (45 degrees at | MHz).

Can a full-amplitude bandlimited input signal such as a square wave rise far enough in 125 nanoseconds to overload any stages preceding the compensation (usually just the input stage)? The integrating action of the compensation pre vents later stages from overloading. A 2V-peak square wave bandlimited to 20 kHz (0.025 (V/uS)/V) will rise about 63 mV in 125 nsec. This is enough to drive some input stages into nonlinearity, but most high-quality amplifier designs have enough local feedback in the input stage to easily handle such a signal.

Remember, this square wave is over three times as fast as the fast program discussed earlier. Feedback factor and open-loop bandwidth are not relevant here, since designs with different values for these could easily have the same in put stage design and delay. The important criterion here is to have an input stage that can handle large input signals, at least under transient conditions.

We must now determine if the amplifier can keep up with the rate of change called for by the input signal without being driven into nonlinearity by internal overshoots. This is basically a question of margin against slew rate limiting, since slewing is the result when the overshoots are so big they are clipped. At this point it is important to emphasize that it is the magnitude of the overshoots which is important, not percentage; many papers in the literature have erred in emphasizing the latter. It should be clear that good margin against slew rate limiting guarantees that the overshoots will not be large enough to cause nonlinearity and thus TIM.

How does feedback factor affect slew rate? By itself, not at all. Let's take an example. Amplifiers with high feed back factors usually have the extra open loop gain after the point of compensation. The most common example is the use of a current-source collector load on the pre-driver stage. Suppose this stage has a shunt capacitor at its input for compensation. If we double the gain after the compensation, we must double the value of the compensating capacitor to restore the gain crossover frequency to its original value. However, the added gain means that we now need only half the rate of change on the capacitor to achieve the same output slew rate, so the magnitude of the overshoot charging the larger capacitor is unchanged. The percentage overshoot is approximately doubled, however, because of the smaller final value at that node. Even though the roll-off starts one octave lower, the input stage doesn't have to work any harder to achieve a given slew rate. Similar arguments reveal the same result for other forms of compensation, such as Miller effect. Similar arguments also show that overload recovery time is not increased by a larger feedback factor.

Amplifiers with large feedback factor and small open-loop bandwidth thus slew just as well and are just as free of TIM as any others built with the same transistor ft's, given proper design. These observations are confirmed by measurements which Matti Otala con ducted on some JFET operational amplifiers which had good slew rate.

They showed extremely low TIM in spite of very high feedback factor and small open-loop bandwidth. Good designs typically require a slew rate margin of" less than 4:1, while a really poor design (gross open-loop nonlinearity, asymmetrical slewing, etc.) might require 10:1. Based on earlier observations, a well-designed 100-watt amplifier should have a slew rate of at least 12 V/uS.

It is not surprising that many of the best-sounding amplifiers have low feed back factors, for they were probably designed by people who were highly aware of, and concerned about, things like TIM. In choosing a low-feedback design, they realized that they had to be infinitely more careful with the open loop design. However, these excellent designs could be further improved with-out penalty by the proper application of more overall feedback.

Finally, it is extremely important to keep in mind that the mechanism being discussed here (and in most TIM literature) is only one of many sources of high-frequency intermodulation distortion (TIM). Adequate slew rate does not guarantee freedom from TIM produced by other sources such as junction capacitance nonlinearity, output transistor charge storage effects, etc.

Very truly yours, Robert R. Cordell Tinton Falls, NJ This is altogether the most clear headed presentation of the subject we've seen so far, and we find it absolutely un exceptionable as far as it goes. What we're not ready to concede-not yet, anyway-is that feedback causes no audible degradation of the signal whatsoever, even if it need not be guilty of slewing induced distortions. The "out-of dateness" of the error-correcting signal presented to the input may be a broader and more troublesome issue than is generally admitted, unless the front-to back delay is truly minuscule, as in a good tube amplifier (e.g., the Futterman). In typical solid-state circuits the delay may be responsible for subtle time modulations. See also our comments on the no-feedback Rappaport AMP-1 in this issue.

-Ed.

The Audio Critic:

After reading your articles on tone arm geometry, I set up the arm I was using at the time, a Grace G-704 unipivot with Supex MC cartridge on a Linn Sondek turntable, and felt that I could hear an improvement. I then tried the same experiment using the G-707 arm in place of the 704 arm with the same cartridge and turntable. Again audible improvement. Not so much between the two arms but between the manufacturer's setup procedure and The Audio Critic's procedure.

Because of the difficulty in carrying out really accurate tests subjectively in a purely domestic environment (I believe it is very easy to subject oneself to a psychological snow job when dealing with audio equipment that you have made a major financial decision over), I wrote to Hi-Fi News magazine in England about your article. Hi-Fi News along with The Audio Critic is a magazine which I have a great deal of respect for, and I had hoped that they might try the setup procedure you outlined and review the result for me. A copy of your article ac companied the request.

Their response was very prompt and, as they suggested, I have forwarded a copy of their comments. The section where they disagree with your argument appears to be quite persuasive and I would appreciate your reaction. I have incidentally repeated my request that they try your setup procedure and let me know the subjective results.

Yours faithfully, Bill Carter Kalamunda, Australia

The letter to Mr. Carter from John Crabbe, the chief Editor of Hi-Fi News & Record Review (Britian's premier Journal for serious audiophiles), follows.

Dear Mr. Carter:

Thank you for your interesting letter of June 8th together with the copy of the article from The Audio Critic. All your points are noted with interest but I must offer a few remarks about the article, since 1 believe it contains one rather serious error.

Although I would not necessarily dispute the claim that quite small errors in tracking angle can be heard (indeed, I have suspected for some time that some of the changes in quality across a disc arise from the changing error with a pivoted arm), I cannot for one moment accept the notion that this is due to a timing error between the two grooves rather than because of old-fashioned harmonic or IM distortion. The article proceeds very sensibly through the argument up to the point where the matter of linear displacement of stylus contact is discussed in detail. Here, it completely loses sight of the fact that such errors as there may be are derived from the final angular displacement between the true radial path and the actual stylus motion path. I must confess that I have not worked through the maths as I am rather busy at present, but I suggest that you make a simple sketch of a groove viewed from above, draw in a line at right angles to represent the correct stylus motion path, and then another line angled away from this by, say, 3° (which should represent the sort of conditions we are talking about). Now, if you put some dimensions in relating to the groove width, you will see that only at the very highest audio frequencies will the contact point on one side of the groove be displaced from that on the other by more than the tiniest fraction of a wavelength. Also, the 5 micron displacement (what they call '5% time modulation') again represents such a ridiculously small fraction of the 100 micron wavelength that one would have to sit with one's head in a rigid clamp exactly equidistant from the two loud speakers before one could begin to consider the audible consequences of such an error.

There is also the point that from about 4 kHz upwards all the evidence on stereo perception suggests that the relative phase of the left and right signals (or their absolute arrival times up to a certain limit) is comparatively unimportant, while of course at lower frequencies the effects under discussion represent left/right differentials dwindling into total insignificance.

The upshot of all this is that while I agree that the errors in question could produce sufficient nonlinear distortion to be audible, I must regard the time thesis as a complete red herring. Incidentally, on the matter of vertical tracking angle, a composite review of cartridges, to appear in our July issue, offers measurements of both harmonic and IM distortion arising from vertical modulation, and finds a significant relationship with the effective vertical tracking angle.

Thank you again for raising this whole topic, and please feel free, if you wish, to send a copy of my remarks to the Editor of The Audio Critic. I would not wish to intervene directly as from one magazine to another, but if you wish to quote me on the subject by all means do so. Of course, you may disagree with my argument and send a string of maths back to me!

Yours sincerely, John Crabbe Editor Hi-Fi News & Record Review Croydon, England

It surprises and even saddens us to see Mr. Crabbe react with casually con descending speculation to something that isn't a trendy new idea advanced by us but has been rigorously worked out by some very distinguished practitioners and widely accepted for many years. In his 37-year-old seminal paper on tracking error, H.G. Baerwald clearly states that the effect of such error is a frequency modulation of the signal by itself, resulting in the generation of sidebands. He also states that the spectral character of these sidebands will create a larger nuisance effect than the harmonics of ordinary amplitude distortion. Although he doesn't use the modern term "'time dispersive," that's what FM is, isn't it? In the stereo era, the equally important work of Bauer, Cooper, Woodward and White all deals with FM-type distortions (specifically FIM and FXM) in connection with vertical tracking error. Thus it isn't our 'time thesis" that's 'a complete red herring" but rather Mr. Crabbe's in explicable rewording of our simplified explanation of the subject as a "timing error between the two grooves." No sir, we weren't talking about the arrival times at the ear from the two stereo speakers but about the audibility of FIM and FXM distortion (i.e., time-dispersive auto-modulations of the signal) in comparison with "old-fashioned harmonic or IM distortion." It is possible that this basic research isn't available in the library or files of Hi-Fi News & Record Review?

-Ed.

The Audio Critic:

I read issue #5 of The Audio Critic with great interest as well as raised eyebrows.

At the end of your Dayton Wright review you state: "Those who feel that the Bryston causes peaks and ringing in the Dayton Wright, whereas the Threshold does not, are invited to send us their revisions of the laws of physics for immediate publication." I have no intention of revising the laws of physics, nor do I intend to defend the Dayton Wright/Threshold (I don't own them and don't intend to buy them).

What I intend to do is to revise your interpretation of the laws of physics. It is an undeniable fact that for every action there is a reaction.

You should be well aware that the Dayton Wright is a very difficult load to drive. Such a difficult load can cause an amplifier to act in a strange manner, in turn causing the speaker to react to its in put.

I have done quite a bit of work in the area of amplifier stability (I have designed a tube power amp as well as a tube preamp for the Win cartridge --neither of these are presently commercially available). By making a simple change in the output stage of my amplifier, thus making it unstable at high frequencies, I can make my little Rogers exhibit headache-producing ringing. This is in complete accordance with the laws of physics. For every action there is a reaction.

Thank you for taking the time to read this letter. I hope that I have shed some light on the subject.

Sincerely, Craig Herberg

Falls Church, VA

Your point is well taken as long as you take our point totally out of context.

Deliberately misunderstanding someone you wish to differ with is an old dialectical ploy.

In context, however, the following givens enter into the discussion: (1) the Bryston 4B was reviewed by us in is sue #4 as an amplifier especially distinguished by its ability to drive the most complex reactive loads without batting an eyelash and (2) the Dayton Wright review characterized the speaker as having specific and discrete frequencies of ringing, several of which coincided with huge amplitude peaks.

To conclude from this that, who knows, the Bryston could after all be un stable and the Dayton Wright okay requires a strong desire to punish us for having made the provocative statement you quote. We refuse to be so punished by someone we believe knows exactly what we meant.

-Ed.

Speaking of deliberately misunderstanding people, at least a little, in order 10 set the scene for interesting dialogues as well as monologues, the master of the genre turns out to be none other than our old friend, Professor Greiner of the University of Wisconsin. Among other things, he has sent us a Xerox of a lengthy communication to Andy Rappaport, in which he nitpicks, corrects and virtually grades Andy's long letter published in our last issue as if it were a term paper by one of his students. He has also bombarded us with numerous marginal notes on our own published comments, the gist of his general attitude being "sez who?"' and "'prove it." In one of his jottings, for example, he feigns innocence of some of the basic writings of Bode on certain aspects of feedback to which we made glancing popular reference and challenges us "to present your 'mathematical' theory, if any, to the public." Come on, professor, why should we paraphrase, print and mail at our expense material which is available in your own engineering-school library? We're a consumer publication, not a university seminar.

Just as a quick sample of Greineriana, we're publishing below a letter written in response to one of our interpolations into Dick Majestic's letter about the RAM 512 power amplifier.

The relevant part of the RAM letter said:

"I can only suggest to you that you try the difference signal check. All that's necessary is to sum the inverted output signal with the input signal after re adjusting for the amplitude difference and look at the resulting difference signal." Whereupon we parenthetically interjected: "Ah! But you have to look at both voltage and current differences!- Ed."

The Audio Critic:

The parenthetical comment marked is nonsense. The input current to a power amplifier is totally irrelevant. Power amplifiers are designed as voltage amplifiers with high and resistive inputs.

The amplifier is made to look like a voltage source to the load and thus it is the voltage transfer ratio that is import ant. It is the voltage transfer ratio that can be easily measured with the technique described in the letter. This happens to be a very excellent and fundamental measuring method.

The output current of a voltage amplifier is determined by the nature of the load and while it is of interest, it is secondary to voltage. Loudspeakers are designed to work from voltage sources.

If the amplifier fails to provide the current called for by the load, the voltage at the output will show this fact.

I am rather concerned that in your quickness to make a comment every little while in these letters, you are not being careful about them. This will eventually damage your creditability.

My high hopes for the technical soundness of The Audio Critic are wavering.

Sincerely, R.A. Greiner Thus is the straw man set up with a bit of deliberate misunderstanding. 'We must speak by the card, or equivocation will undo us," says the Bard.

Professor Greiner is implying that we need to be convinced of Ohm's law, namely that I=E/R. We already believe that, but while he is looking for evidence of academic delinquency we're looking for good, practical bench tests of amplifiers.

With transient test signals and real-world reactive loads capable of energy storage.

(Like, for instance, loudspeakers.) And under those circumstances it happens to be quite practical on occasion to bring out the current probe. It just might show something about the behavior of the amplifier that wasn't obvious with the voltage probe. Not because Ohm's law doesn't work but because of the way non linear reactive loads work.

We therefore feel that 'nonsense' and 'wavering ... hopes for the technical soundness of The Audio Critic" are somewhat unfriendly. Not to mention "creditability," which is not only un friendly but also a solecism. Why, oh why must one of our longtime academic heroes send us pictures of his feet of clay?

-Ed.

The Audio Critic:

I have just finished watching a very interesting and enlightening show on educational TV which brought one up-to date on computers. Most pertinent was a computer which is able to diagnose over 800 diseases from over 2500 different symptoms. The abilities of computers to analyze, design, and even assemble seem almost limitless. Considering your observations about speakers and their design, and about your correlations between tests and your own subjective likes, it seems to me that it would be possible to feed all the necessary data into the proper computer and come up with an *'ideal" speaker design-which could include everything from size, number and kinds of drivers, crossover networks, proper time/space coherence, etc., etc. I do realize that computers already play a very large part in the production and design of audio equipment, but with your own findings and the right contacts in the computer world you fellows should be able to come up with a most interesting, if not incredible-sounding, speaker system.

What do you say?

Sincerely, Stewart Glick; Rochester, NY

We refer you to the article by Bruce Zayde in our second issue (Vol. 1, No. 2, pp. 28-31), in which he states that all the mathematical linkages between the relevant parameters of low-frequency speaker design can be programmed into a computer and correct solutions extracted at will. This is not true for the higher frequencies; the mathematical models involve extremely complex problems in membrane physics, and the equations are monstrous beyond belief. To the best of our knowledge, complete mathematical models of full-range speaker systems and computer programs for designing full range systems based on such models have not been devised yet. We have seen a program for woofer designs and can report that it works very nicely.

Computer-aided design is becoming fairly common in the speaker industry, but it isn't applicable across the board. Need less to add, computers don't have the answers to fundamentally unsolved problems; they can only supply existing answers faster.

On the other hand, regardless of the possibilities of computerized design in formation, The Audio Critic is definitely interested in putting together a better system than anything available today at any price. (See also the article on reference systems in this issue.) We may end up having to fabricate at least a few parts of such a speaker ourselves, though.

And we're virtually certain that the only moving-coil driver in it, if any, will be the woofer. The rest will be strictly force over-area. That much we know at this juncture even without computers.

-Ed.

The Audio Critic:

I suppose that as many people write you about what shape or format is best for your magazine as tell me at the end of a concert or workshop what aspect of my career I should pursue.

I find one lack of consistency in the pursuit of your goals. The problem is that your definition of advertising is a narrow one. That is that you accept no advertising by "manufacturers ... or other commercial interests." Fine, no criticism of this policy. Yet in the loose sense a published letter on a technical matter, or indeed philosophical as in the case of An drew Rappaport's letter redefining a way of looking at distortion, is, even without your commendation ("Thank you, Andy 7), 100% advertisement, albeit general and unsolicited. You must realize that at least subliminally, if not overtly, you are giving us not only objective reviews but the feeling that we can wait and expect something great to happen (which hasn't really happened yet).

Your review is obviously Gordian knotted on the Rappaport subject.

How can you say that "it did extremely well, giving all comers a hard time in A-B tests" and say also that there is "a barely perceptible thickness . . . in the midrange"? That it "has become ultra smooth and listenable in the upper octaves; we prefer it even to the Hegeman in that respect" but yet that it was "squeezed out by the new generation"? So you essentially gave the designer top billing and the product confused billing.

Are you taking credit with these mealy competitive adjectives for inspiring a mature 20-year-old to run faster to clobber Mark Levinson? In sum I find that letters speak for themselves without your lucid interjections, and the credibility of your stating "no advertising" would be in no jeopardy if your irrelevant congratulations or antic endorsement of an idea were omitted. This would make for more cohesive and powerful editorial ship. Comment after an article-don't interrupt even a jackass's train of thought. They and you speak for them selves.

Best Wishes, Steven Silverstein

Player-Instructor-Designer of Wind Instruments

Stony Point, NY

You've got it all mixed up, Steven.

We accept no advertising, not because the commercial message might influence you but because the money might in fluence us. We thought that was clear to every one of our subscribers, but it appears from your letter that it isn't. We want to influence you profoundly and we'll make every editorial attempt to make you see things the way we do.

That's what critical reviewing is about. If we believe that Andy Rappaport is a very bright young man, or that some other audio designer is an orifice where the sun never shines, we'll make an effort to imbue our subscribers with the same belief.

And that has nothing to do with impartiality in equipment testing. Should Andy come up with a lousy design, we'll say so-and he knows it. On the other hand, if he paid us at a certain rate per page or half-page for advertising . . . Get the picture? As far as conflicting qualities in a piece of equipment are concerned, we can't help it. That's the way it is. If a single preamplifier had the best highs, midrange and lows, the lowest noise, the clearest detail, the best construction, etc., etc., there'd be no problem. In this imperfect world, however, we have to point out the trade-offs and then filter our top choice through our own structure of values.

Your criticism boils down to two things: (1) that The Audio Critic has strong opinions, which it injects wherever it can and (2) that even this opinionated critic hesitates sometimes when choosing the "winner" in topflight competition. On both counts, you're absolutely right.

-Ed.

The rest is back talk from manufacturers in response to reviews in the last issue.

The Audio Critic:

It is an unusual task to respond to three reviews at once, particularly when one review is quite favorable and the others quite the opposite. But here goes.

First of all, thank you for the review of Sleeping Beauty Shibata. Your fin dings concur with those of many other critics. Perhaps if you had auditioned it through one of our Goliath II moving coil cartridge preamplifiers you would have liked it even more.

Now to Grandson. It is difficult enough to understand the subjective terminology used to describe the sound of components without dealing with such subjectivity where it is not necessary. I am referring to such expressions as "(Grandson shows) an unhealthy amount of ringing with capacitive loads." How much is unhealthy? 5% overshoot? 10%? All amplifiers which use feedback will ring into capacitive loads. Grandson does not ring any more than most other amplifiers; in fact, it rings less than most.

Grandson is certainly not current deficient. To quote from your review of the ML-2 in the same issue, "You can't call any amplifier with a 10-ampere peak current capability a little amplifier."

Grandson has 11 amps peak. That aside, the fact that Grandson is able to double its 8-ohm power output into 4 ohms, and triple it into 2 ohms, indicates a sufficient power supply. Grandson uses the identical power supply as Son of Ampzilla, with a lower voltage transformer.

Again, back to subjective terminology when it is unnecessary. What does it mean that you found Grandson's distortion figures ''far from impressive'? What exactly does "some peculiar cross coupling effects" mean? You shouldn't be afraid to quote your measurements.

Perhaps if you had, we could discover that your sample of Grandson was defective, or had some minor problem. Or that it was behaving normally. As it stands now, we have no way of knowing what exactly caused the problems you experienced.

As far as Thalia is concerned, your opinions of its sonic qualities run contrary to our experience, and the experience of our customers. We will simply state that the unpleasant sound you report cannot be attributed to a properly functioning Thalia (or Grand son).

For proof of this we can only ask that readers of The Audio Critic audition these products themselves in the best acoustical environment available.

Readers who are interested in reviews of Thalia and Grandson which come to opposite conclusions from those of The Audio Critic are invited to write to GAS.

Sincerely, Adam Zareba

Vice President, Sales & Marketing GAS Chatsworth, CA

(1) Even if Goliath II is the perfect moving-coil pre-preamplifier (which we seriously doubt), it can't be as good as an equally perfect transformer. (See Vol. 1, No. 5, p. 57.) It's quite possibly better than a mediocre transformer, but that's not the kind we use.

(2) Grandson does ring more than most amplifiers. The overshoot into a 2 microfarad capacitor across the load resistor is at least 100% and the ringing is undamped. What's more, a similar signal of lower amplitude appears at the output of the undriven channel. Is that the non subjective terminology you were looking for? (3) Maybe Grandson isn't current deficient in the grossest sense, but it's certainly a lot cleaner into 8 ohms than into 4 or 2 ohms.

(4) We did give a specific example of Grandson's distortion figures.

(5) If either the Grandson or the Thalia hadn't been properly functioning, you would have been able to determine that when the units were returned to you.

-Ed.

The Audio Critic:

In regard to the model 3100 pre amp report in The Audio Critic #5, we wish to make the following comment.

We hope that it will be run unedited in the next issue.

The results obtained by The Audio Critic on the 3100 Ace Audio preamp do not agree with those obtained in other magazines, where the reviews were done by other veteran reviewers of many years experience. If any of your readers are interested in seeing these reviews, we will be pleased to send copies upon request.

A number of our customers, or prospective customers, have called to tell us that they have compared the 3100 or 3000 preamps to very expensive pre amps costing over $1000 and some lesser ones, and have been unable to tell which is playing. These tests have been done in the comfort and leisure of their homes, under relaxed and intimate listening conditions with familiar records, etc.

In view of the discrepancies obtained in your test, we are requesting an immediate retest under carefully con trolled conditions, and we hope that we may be allowed to participate in some of the sessions.

Thank you.

Yours truly, John Grauer President Ace Audio Co.

East Northport, NY The main reason why many thousands of audiophiles subscribe to The Audio Critic is that the recommendations of "'other veteran reviewers of many years experience' are un reliable. If our findings duplicated those of other publications, there would be no need for our services. Specifically, in the case of preamplifiers, reviewers who don't align their reference cartridge and arm for optimum lateral and vertical tracking geometry can't possibly come to a valid conclusion about the sound of a phono preamp. The same goes for "customers . . . in the comfort and leisure of their homes." As for a 'retest under carefully controlled conditions," we would first like to be told what was wrong with the conditions specified on pages 49 and 50 of our last issue. Were they carelessly controlled or was their main fault that they didn't result in a higher ranking for the Ace 3100? One more thing. We're curious to know why John Grauer considers his letter so devastating that he feared we might run it edited.

-Ed.

We saved the rarest gem of our collection for the last. Judging from the salutation, it may also have been sent to, or even published by, others.

An Open Letter to Peter Aczel, Editor and Publisher of The Audio Critic (Enquirer):

In reference to your review of the Hartley 24" subwoofer, which appeared in your magazine, Volume 1, Number 5, I want to state for the record that it is highly unethical for any magazine to review a product listing a manufacturer, when in fact that product in review was constructed by someone else. This, of course, is the case ... the subwoofer system you tested was not manufactured by Hartley Products Corporation. I have also noted that in almost every other re view in your last two issues, the product was obtained from the manufacturer or authorized distributor.

Question--Why not ours? Although this fact would end your review before it began, I feel I must comment on your rather pompous approach approximating a review.

The Lab Measurements:

I have noted your laboratory instruments described in Volume 1, Number 4.

Although your equipment is not significantly different from any good warranty station, I am amused that you see your self as the "Audio Messiah", able to comprehend all mysteries behind frequency and time. After all, you said ". . . even so, we're convinced that if every speaker manufacturer routinely performed our simple tests (as we know for a fact nine out of ten don't), speaker design would be a far less haphazard affair today and the art would advance more sure-footedly." Since there are, at last count, approximately 200 nationally branded speaker companies, [ would like to receive your list of those 180 speaker manufacturers that don't follow Peter Aczel's easy method of measuring loud speakers.

The Listening Test:

For the record, you state in your masthead that The Audio Critic represents the personal findings and judgments of the editor and staff. Before measuring you (and presumably, your

= . even so, we're convinced that if every speaker manufacturer routinely per-ance was. [ quote, "We suspect that the vast majority of audio practitioners, whether on the manufacturing or the reviewing end, have never really experienced the startling clarity of a system with optimized time-domain characteristics from stylus tip to speaker diaphragm. Seasoned audio people who visit our sound room are invariably astonished, although we don't do anything there that they themselves couldn't duplicate if they ordered their priorities as we do." As our newly anointed Messiah, we think you owe it to struggling audiophiles everywhere how we, too, can reach this nirvana. By the way, a list of your disciples might help to enlighten us as well.

The Admonitor:

I quote, "The best woofer cones are soft, pulpy and lossy. 'Piston-like motion' at, say, 100 Hz can be obtained with a woofer made of mucus; it has nothing to do with 'structual strength' but with the ratio of the cone radius to the wave length." Mr. Aczel, you will be happy to know that we now have, on our drawing board, a radically new woofer design.

The diaphragm (which is 36' in diameter) is entirely constructed from the mucus membrane extracted from the Great American Wild Boar. Should we submit this for review? Final Comment:

I would like. to encourage all audiophiles who have read the review (sic) and who are just as outraged as we are, to write us with the intent to learn about the products we spent 50 years in designing.

I sincerely believe that the advice given by Peter Aczel and staff in The Audio Critic is an evident dis-service to consumers and has no place among other audiophile-oriented publications.

Richard Schmetterer

President Hartley Products Corporation

Ramsey, NJ

Since Richard Schmetterer obviously fails to offer a single technical argument to repudiate our findings, and since his attempts at derision and insult are so witless as to be totally without sting, we shall leave this pathetic little tantrum unanswered except for the first paragraph. That at least makes a point, albeit a totally misleading one.

Hartley Products Corporation does indeed manufacture the 24-inch driver in the subwoofer system we tested; in fact they're selling it routinely and without any protest as a raw driver. What they don't manufacture is the enclosure in which we tested this driver. They do, however, offer a full-range ''Master Reference" speaker system incorporating the same woofer in a sealed en closure of approximately 18 cubic feet internal volume. The custom enclosure we used was also sealed, also of approximately 18 cubic feet internal volume, and fanatically well-made and braced. Therefore, what Richard Schmetterer is implying is that his 18 cubic feet of New Jersey air would have drastically lowered the Q the woofer exhibited in 18 cubic feet of New York air.

If that's the case, that magic air should be bottled and sold to speaker designers in six-packs.

As our subscribers are well aware by now, we obtain equipment for reviewing any way we can. Borrowing it on a 90 day or 180-day memo from the manufacturer or a distributor is the most convenient; borrowing it for a few weeks from a dealer or a friend is a little less so; buying it ourselves is the last resort but we've done it. Furthermore, we document the provenance of each item at the beginning of the review. So we didn't single out Richard Schmetterer for a special handicap. He's doing a pretty good job of it on his own.

-Ed.

---------

[adapted from TAC]

---------

Also see:

The Audio Critic Seminar on the State of the Art: Part II

Why We're So Mean, Vindictive, Arrogant, Negative--and Truthful

Various audio and high-fidelity magazines