LessLoss's Louis Motek elaboration
Every time I write about an audio engineering solution of ours, I must first consider the adressee, the one that the writing is intended for. Is it going to be an artfully minded individual who has never studied circuitry, who is more interested in subtle auditory nuances he'll experience listening to music, who wants to understand in what general way our gear isto be differentiated from other solutions; or, is it going to be an engineer, who is interested in a dry technical explanation, perhaps even with accompanying formulas and detailed physical calculations?
Most often, I write addressing the former, aiming for a popular style, which is hopefully understandable to anyone who does not have a formal electrical engineer's education.
Sure, this can cause suspicion to readers who have a background in engineering, who suspect logical cut corners. So, this time, I will write in an engineering spirit, trying to answer some frequently asked questions, some of which were raised by your readers.
Why, in a power cable, does one need to filter Radio Frequency (RF) signals, if we cannot hear them anyway?
Considered from a Radio Frequency (>100 MHz) standpoint, any piece of audio gear, be it an amplifier or a player, is a nonlinear converter. Not only with regard to its input, but alsoto other signals which enter it from without, meaning directly „from the air" in the form of EM waves, and through the RF garbage picked up by the power cable.
In other words, our gear is, more or less, a radio receiver, or, more precisely, a demodulator.
We have often heard a variety of undesired sounds from our loudspeakers. For example, when our mobile phones receive a call, when the ringer is turned off, or unexpected sounds from nearby radio stations, even though we have no radio receiver set up in our systems.
In audio electronics, we are concerned mainly with four types of demodulation:
1. Amplitude demodulation. This is when the interference, in the form of a RadioFrequency undergoing a change in its amplitude, is directly demodulated, meaning that the change in amplitude corresponds to a change in a frequency low enough for us to hear. Any nonlinear element in the circuitry of our gear runs a chance of behaving as such a demodulator: a diode, a transistor, or a vacuum tube. Even a rusty bolt! (http://en.wikipedia.org/wiki/Rusty_bolt_effect ) When it is loud enough, it is recognized asradio interference. This type of demodulation is not held to be particularly surreptitious or deceptive because it occurs independently from the desired signal. Since it is not correlated to the desired signal, this type of interference is interpreted by the listener as any other extraneous acoustical noise.
2. Phase or frequency demodulation. This is when an RF interference's phase or frequency change is extracted, which can occur at audible frequencies. In our audio gear, any nonlinear selective circuit can become this type of parasitic frequency demodulator. The largest chance for such a circuit to be formed is in a phono correction unit, which is part of why these are often the most sensitive devices in audio systems.
3. Desired and undesired signals' amplitude intermodulation. This is a more deceptive type of demodulation, occuring when two or more signals' harmonics interact. It is deceptive because it can be entirely inaudible while there is no desired signal (in the pauses), but appears as a form of sound distortion while the desired signal is in playback.
4. Digital players' or digital-to-analogue converters' digital clock frequency modulation, akaJitter. This is the most deceptive case, since the human ear is incredibly sensitive to incorrect phase reproduction or phase distortion. Before the dawn of the digital age, we were pestered by magnetic tape „stretching", a type of vibration perhaps most famously audible in piano recordings. In digital players, Jitter is also a type of vibration, with the difference being that its spectrum can be very wide. One reason Jitter drives engineers up the wall is that measuring it directly with equipment is not straight-forward. Because the speeds are so high and the levels of useful data so low, very expensive measuring equipment is necessary, and the art of measuring Jitter is open to much interpretation. Setting up a test can be very difficult, and results can vary from engineer to engineer. But even in small amounts, Jitter can be heard readily. Throughout this battlefield between the engineers and the audiophiles, many arrows have been seen to fly. Jitter's ability to be heard has not only to do with its amplitude, but also with its spectrum, leading to the even more complicating affair that Jitter at a lower amplitude can be even more audible than Jitter at a higher amplitude, if its spectrum is different. Spectra can come in infinite varieties, yet further mudying the waters.
How can RF interference cause Jitter?
Let us first recall that digital players or DAC converters often contain more than one clockoscillator, which can be run by a quartz or a PLL (phase locked loop).
In general these are oscillators which are made up of quartz crystal or energy-storingcircuits and active comparator circuits, which are usually transistor-based.
Interference, upon entering the clock's circuitry through the power supply chain to the active circuit, changes the comparator's sensitive threshold level, which results in an erroneous adjustment of the following clock signal's front either earlier or later in time.
High-end audio gear manufacturers go to incredible lengths in their attempt to keep Jitter to a minimum. Jitter does not appear only in the clock oscillator. It grows and becomes more complex at every clock transition in the circuit, for example in the SPDIF cable, and in the SPDIF signal receiver circuit, where the clock signal is often transmitted together with the digital data signal. Interference also can enter the SPDIF signal directly and cause Jitter. Hence the importance of good shielding.
Why place importance on the power line's „last six feet" of wire? How can the last 6 feet of audiophile power cord affect the sound, when the electricity has alreadytravelled many miles through just standard copper and aluminum wires?
There was a time when this question gave cause for my own share of sleepless nights, too.But I believe it is possible to explain, and I don't feel we must delve into „esoteric" theoriesin order to do it properly, either. Let's take a look at how our gear appears to an RF wave:
Power cords constitute an antenna system which absorbs local RF waves and feeds these into the equipment case, the ground and the power supplies wires.
Our job is to ensure that the power cable induces the least possible amount of electromagnetic waves, and filters radio frequency interference which already exists in the power line at the wall socket.
Truth be told, because power lines are not created for RF signal transmission, high frequencies do not travel down them very far.
Only longwave signals stand a chance to travel down them a long way, and these are efectively filtered out by RC or LC fitlers which are located within the power supplies of all audio gear.
At the very last few feet of the power line (which represents the power cable itself), the most damaging frequencies are received and fed into our sound system, because these frequencies are hardly filtered by traditional high frequency filters.
The formula is:
F = 300 000 000 (m/s) / (length of antenna in meters x4)
When the cable is 1 foot long, or 0.3m, then f = 250 MHZ
When the cable is 1 m long, then f = 75 MHz
When the cable is 2m long, or about 6 feet, then f = 38 MHz
Why don't we use traditional LC filters („caps and coils") for power filtration?
The main reason is that they are not effective. Most often, such a filter is effective at filtering only a narrow spectrum.
Furthermore, such a filter must be connected with the gear by means of a cable anyway, which will become a perfectly good high frequency antenna.
Mono & Stereo high end audio magazine
All rights reserved