Stop Digital Madness! Page 3

The only cyclical information above 25kHz that's part of any vinyl recording of music is the continuous 40kHz tone that carries the difference information on RCA's CD-4 quadrophonic recordings. By the time this 40-kHz carrier is run through RIAA equalization, it is no more powerful than the rest of the musical signal (footnote 4).

But let's grant her the benefit of the doubt for a moment and pretend that all digitally mastered source material does contain the pulses she sees as the source of digital damage. Let's follow the level of those pulses through a topnotch audio system.

In the worst case, of an MC cartridge generating 1mV output at 1kHz, there could be 300mV pulses of 0.016ms duration when mistracking. These are very strong in comparison with the rest of the audio signal, but 300mV is one fifth of what you get from a flashlight dry cell, with nowhere near the dry cell's available power. There's no way a flashlight battery can crack a platter bearing unless it's used as a blunt object.

Now let's assume that those 300mV pulses don't just overload the pre-preamp and get clipped off, which is what usually happens. Assume that they're amplified along with the signal. Before the program level reaches 1 volt, the signal will go through the RIAA equalizer, which will knock about 30dB level off the pulses, putting them at +20dB relative to the main 1 volt signal, or 10 volts above it. This is pretty big by comparison with 1 volt, but in order to even make a mark on a platter bearing with 10 volts, you would have to deliver it to the bearing via wires, and touch them directly to the bearing surface. But we're not doing that. The bearing is safe inside its steel or brass well, the whole assembly is at ground potential (or floating, if you didn't ground your turntable), and the audio signal is enclosed in the metal preamp chassis, which should prevent the escape of even the miniscule electrostatic fields the pulses produce around the wires carrying them.

By the time they get to the power amplifier, the ultrasonic pulses may approach in intensity the amplifier's maximum output capability (they cannot exceed it). Now, here's a possible source of destructive energy! A frequency of 60kHz is considered to be a Radio Frequency (the bottom of the AM radio band is 50kHz), and 200 watts is only a bit less than half the power at which commercial AM stations operate. But a radio antenna delivers most of its power into the air; an amplifier delivers most of its power to the loudspeaker. And the parallel construction of speaker cable acts to cancel most of the signal that might otherwise be radiated into the air. But it's obvious that some spurious RF is going to be floating around in your listening room, but I seriously doubt it could damage a turntable?

I'll leave out the question of turntable damage from the acoustic waves radiated by the loudspeaker or the electromagnetic radiation from speaker cables. To the best of my knowledge, speaker hookup is unnecessary to Prof. Reilly's findings. Even if it were, the quantities of energy available acoustically or electromagnetically are truly minute.

My real question concerns how much power is required to induce micro-cracking in turntable spindles. Ultrasonic energy in the range of 1 watt and below is routinely used in industry for the nondestructive checking of the structural integrity of metal castings, but it takes at least several watts to crack steel under ideal circumstances, using the object's natural resonant frequency.

Unfortunately, I see no way to get enough ultrasonic energy to the turntable spindle or bearings to induce microcracking. 80mV is less than one-tenth the level used for nondestructive testing, and none of the devices in an audio system carrying higher energies have a way of delivering that energy to the turntable! If there is no way of getting enough ultrasonic energy to those bearings to crack them, how do we explain Prof. Reilly's findings? We don't. And I have a strong suspicion we won't have to, because, very frankly, I find this business just a little bit fishy.

One high-end turntable manufacturer of our acquaintance has already attempted to duplicate her findings, and has found no correlations between the kind of disc played and the speed variation of his 'tables. Moreover, the kinds of speed variations discovered by Professor Reilly—up to 1.5%, alternating over a 10-minute period between +1.5 and -1.5—are gross; the turntable manufacturer was unable to find anything approaching that much speed variation, and the off-speed conditions he did detect were stable. All this without respect to the type of discs played. I suspect others will have no better luck corroborating Prof. Reilly's findings.

And I am suspicious of Prof. Reilly's use of the word "digitized" in referring to turntables which have ostensibly been damaged. As a scientist, she should know better than to so abuse a term whose meaning is clearly defined. To "digitize" means to convert into binary code. No one has yet done that to a turntable platter.

I am suspicious about that claim of 100% consistency in Prof. Reilly's test results. Though it isn't impossible, 100% consistency in any experiment is very unlikely.

I am also suspicious of Neil Levenson's oscilloscope traces, which provide the "evidence" of ultrasonic pulses. On those traces, the pulses taper in width, as though accompanied by strong sidebands, and they are almost symmetrical, having similar positive and negative amplitude. In other words, they are not "pulses," but very brief bursts of noise spanning several kilohertz. They do not look like the mistracking pulses that others have observed, nor do they resemble anything involved in the digital audio process.

I may later regret my acute skepticism. We have a department in Stereophile for expressions of such regret, called "Crow Pie." As of now, however, I predict that one of three outcomes will follow Professor Reilly's pronouncement, though not necessarily in this order of probability:

(1) She will find her measurements to have been faulty, and will send a guest submission to "Crow Pie;"

(2) No one will be able to duplicate her findings, relegating them to "curiosity" status; or

(3) It will be proven beyond a doubt that Professor Reilly's distaste for Ones and Zeroes so stimulated her latent and hitherto-undiscovered psionic ability as to interfere with the normal operation of the turntables she tested. She will then become an international celebrity.—J.Gordon Holt

Footnote 4: At 6dB/octave, an RIAA phono section reduces HF energy (relative to 1kHz) by about 14dB at 10kHz, 20dB at 20kHz, 26dB at 40kHz, and 32dB at 80kHz. There are two other ways an ultrasonic tone can get onto a disc: intentionally, in the production of a frequency test record, and inadvertently, when an improperly oriented cutting stylus chatters rather than cutting smoothly through the nitrate surface of the master.—J.Gordon Holt