New IHF standards for amplifiers

The Institute of High Fidelity Announces:
NEW STANDARD FOR AMPLIFIERS

NEW MEASUREMENTS AND REVISED TEST TECHNIQUES PROVIDE
MEANINGFUL COMPARISON DATA FOR THE AUDIOPHILE AND ENGINEER
By Daniel von Recklinghausen

IT does not ordinarily come to the consciousness of the hi-fi enthusiast that most of the equipment he buys, whether receiver, amplifier, or turntable, is manufactured to meet a rigid set of specifications, or standards, which, taken together, constitute the level of performance these components must attain. These standards are either the manufacturer's own or those agreed upon with other manufacturers through their professional association, the Institute of High Fidelity. The existence of standards is important not only to the manufacturer of hi-fi equipment, but to the buyer as well, because they provide a common basis of discussion and comparison.

In the case of high-fidelity amplifiers, the Institute standard since 1959 has been its IHFM-A-200, but the Institute's Amplifier Standards Committee has now finished work on a revised standard: the IHF Standard Methods of Measurement for Amplifiers. The new Standard is an extension and expansion of the old one that it replaces, adding instructional material, measurements for stereo amplifiers, and further tests of amplifier characteristics that laboratory and manufacturing experience have shown to be of importance. The new Standard has been framed to include not only tube and transistor amplifiers, but is also purposely phrased so that any other amplifying devices that may come along in the future can be accommodated under its terms.

It is not possible within the space of one short article to describe and to explain in detail the entire new amplifier Standard. (The Standard itself is perhaps 10,000 words in length, and any detailed explanation could easily be three or four times as long.) However, some of its salient aspects are worth discussing both from an information point of view and because they illustrate the IHF Standards Committee's approach to the entire question of standards.

Over the years, it has become increasingly evident to IHF members that the previous amplifier Standard was inadequate for two principal reasons: (1) two amplifiers could test the same, but sound radically different; and (2) additional specification parameters were needed to provide goals for the engineer working for design improvements. The solution to both these problems is a more comprehensive set of tests and measurements, and this is what the new Standard provides. For example, the old Standard specified that eleven different aspects of mono amplifier performance were to be measured and given numerical values. Under the new Standard, on the other hand, nineteen different numerical ratings, plus a total of thirty-one different graphs are established for the complete measurements of an amplifier. To spend the time necessary to make all these measurements on every amplifier would of course be impractical. This problem is recognized-and solved-in the new Standard by making most of these readings and graphs optional. The complete set of measurements thus provides ample information for advance design guidance, and as few as seven of the most important of them are ample for purposes of specification and manufacturing. The seven minimum amplifier specifications are listed in Table I, the balance in Table II.

One of the most important characteristics of an amplifier is its output power. In the immediate past, observers of the audio scene have noticed the development of a strange situation in which the same stereo amplifier might be rated-depending upon the whim of the manufacturer-at anywhere from 15 to over 100 watts. This situation came about for a variety of reasons, both commercial and technical, but in any case, new and firm standards dealing with power and the distortion level at which it is measured were obviously needed. Even the Electronics Industries Association (EIA), the trade organization of the radio and television manufacturers, became convinced of the necessity of a standard rating method and therefore established its own amplifier power standard for EIA members specifying that equipment power ratings be taken at a harmonic distortion level of 5 per cent. But for high-quality music reproduction, this 5 per cent figure is much too high. Hi-fi component manufacturers rate their equipment at distortion figures ranging from under 1 per cent up to a maximum of 2 per cent.

However, because there is still a lack of agreement among hi-fi manufacturers as to the most appropriate distortion level at which to rate an amplifier, neither the old nor the new Standard specifies a particular distortion figure at which power is measured, and it is thus left to the option of the individual manufacturer. Each manufacturer has therefore chosen what he believes to be the optimum figure for his own equipment. For example, if he chooses to rate his amplifier's power at a very low reference distortion level (say, 0.6 per cent), then the rated power output will be somewhat lower and the power bandwidth will be narrower. (The relationship between rated power and rated distortion at several distortion levels is illustrated by the power-bandwidth graph, Figure 1.)

Under the new Standard, three steps are necessary to arrive at a power rating: (1) power with respect to distortion is measured; (2) a curve is drawn; and (3) the curve is analysed to provide a verbal statement of the required data. To do this, the manufacturer of the amplifier decides at what distortion level and at what power output he wants his amplifier to be rated. These two manufacturer-chosen reference characteristics for a particular amplifier are shown as dashed lines on the graph in Figure 2. For purposes of discussion, let us say that the manufacturer has rated his amplifier at 25 watts at 0.6 per cent total distortion. The point on the graph at which the output-power curve derived from steps (I) and (2) above crosses the horizontal reference line of 0.6 per cent distortion is the amplifier's rated output power. As stated in (3) above, the verbal statement of the specifications is then derived from the graph. In the example of Figure 2 (solid curve), the amplifier has bettered the manufacturer's specifications in that at the manufacturer's rated reference power (25 watts) the amplifier has 0.25 per cent instead of 0.6 per cent distortion and at the manufacturer's rated reference distortion (0.6 per cent) the amplifier is capable of producing 35 instead of 25 watts output. The IHF Standard requires that, in testing, both these figures be given-the total distortion at the rated reference power, and the maximum power at the rated reference distortion.

In some amplifiers, this crossing of the reference distortion line may take place not only at some high output power point, but a rise in distortion may also occur at a low output power (see the dashed curve in Figure 2). The new IHF Amplifier Standard requires the listing of this increased percentage of distortion (1.15 per cent) and the power (1.8 watts) at which it occurs (points A), and also the two values of power (0.4 watt and 6 watts-indicated as points B) where the reference distortion line is crossed. This allows a manufacturer or test lab to make a formal distinction between two amplifiers: one that has a rising distortion at low-power levels, and one that does not. Insofar as the manufacturers and test labs make these new figures available, the buyer is in a much better position to choose between two amplifiers.

The new amplifier standard also specifies, as part of its definitions of amplifier characteristics, the nature of the test equipment to be used. For example, distortion is defined as the reading of an instrument that indicates the total residual hum, noise, and distortion components between 20 and 200,000 Hz. (Hertz, or Hz, is the new term for "cycles per second" recently adopted by the U. S. Bureau of Standards and rapidly coming into general use.) Therefore, the distortion meter responds not only to total harmonic distortion in the amplifier's output signal, but also to modulation distortion, oscillation, hum and noise, and everything not a part of the pure sine-wave test-signal input.

"Power" itself is also defined in the new Standard, and the various ways in which amplifier power is described are recognized. There is continuous power, which an amplifier should be capable of delivering for at least 30 seconds, quite long enough to make a measurement and also long enough that any power-supply instabilities within the amplifier will have disappeared. The measurement is made individually, one channel at a time, and also with all channels operating simultaneously. (Reference is made in the Standard to "all channels" instead of "both channels" in order not to exclude future amplifiers that may have more than two.)

Of course, audio amplifiers are used in the home not for the reproduction of sine waves, but for the reproduction of music, speech, and other program material. And unlike a sine-wave test signal, program material varies constantly in amplitude. Almost every amplifier can produce a higher power output for a short period of time than it can for a long period of time-say 30 seconds. Audio engineers also know that an amplifier may possibly test well on sine waves, but then, in normal operation with program material, generate low-frequency transients and other forms of instability and distortion. The old IHF Amplifier Standard recognized only that an amplifier could produce more power while playing music, and therefore set up a "music-power" measurement by assuming that the amplifier's power-supply voltages remained constant under the short-term power demands of normal program material. The measurement of music power therefore involved maintaining all the supply voltages within the amplifier at the same values as they would be with no signal going through the amplifier and then making power and distortion measurements at leisure. It was felt by the IHF Standards Committee that this measurement in itself was neither sufficient nor meaningful. The Committee therefore prescribed that a second measurement should also be made using a special switched sine-wave test signal whose waveform build-up resembles the attack characteristics of music and speech. Output power and distortion measurements are made during this "turn-on" period of only a hundredth of a second by analysing the waveform on a calibrated oscilloscope. This measurement not only shows up as distortion whatever harmonics, modulation products, or noise the amplifier produces, but also indicates any transient instabilities in the amplifier that might appear with a music signal, but not with a test signal.

According to the terms of the new Standard, both of these types of measurement-the older music-power and the new transient-distortion tests-are made and curves showing the relationship between output power and distortion are drawn. The curve yielding the lower power (or higher distortion)-in other words, the "worst" curve-is used for the dynamic-power rating of the amplifier, replacing the older music-power rating. In a stereo amplifier the two channels are measured separately with a signal applied to only one channel for single-channel performance, and a signal applied to all channels simultaneously for multi-channel performance. Measurements performed in the author's laboratory and elsewhere have shown that this new dynamic-power measurement technique is quite effective, in that it provides a far closer correlation between amplifier measurements and listening quality than was possible under the old standard.

Power, of course, is only one of .the many performance aspects of an amplifier; even the old Standard included such important measurements as power bandwidth, sensitivity, frequency response, and signal-to-noise ratio. The new standard specifies all these measurements and includes performance of controls, interaction between controls, and so forth. In addition, such other important information as input impedance, output impedance, and amplifier stability must be supplied.

The new Standard will help establish design goals for audio engineers and at the same time furnish test techniques for validating them. For the audiophile, the new ratings will make possible a more intelligent choice among the profusion of amplifiers now available. Copies of the new IHP Standard Methods of Measurement for Amplifiers can be obtained from the Institute of High Fidelity, 516 Fifth Avenue, New York, N.Y. 10036. Price: $2.00.