Induced Electromechanical Distortion (continued)
We have all heard transformers "hum" - right? In most cases they are being powered by a 60Hz AC utility power line (wall outlet, etc.), but what you usually hear coming from them sure doesn't sound like a clean, pure 60Hz tone. There ya go - The "noise" they generate is a result of the 60Hz signal feeding or "driving" them, so we know the two are related,... but the resulting spectral content of that noise is anything but an accurate reproduction of the original 60Hz signal. Hmm.
The Electrostriction Effect:
The Electrostriction Effect is caused by AC signals (like music, 60Hz AC Power, test tones, etc.), and when applied to a typical electrical capacitor it will cause the capacitor plates to be attracted to each other with every half-cycle of the applied signal. Then the reverse force called the " Piezoelectric Effect" causes the same capacitor to generate a related AC voltage/signal, but the frequency of this newly generated signal will now be twice that of the applied signal and delayed in time with respect to it. Therefore, the capacitor has effectively become a generator of phase-shifted 2nd Harmonic Distortion, and possibly multiples thereof which. In the case that these harmonics happened to be even-ordered multiples (2x, 4x, 6x, etc.) one might be tempted to be less concerned about the matter. That's because even-ordered harmonics are often considered to be relatively unoffensive sounding and sometimes even pleasant, such as in the case of many tube-based amplifiers or even musical instruments.
Spectral analysis of the resulting capacitor signal though would likely reveal it to be comprised of far more than just simple 2nd and/or even-ordered Harmonic Distortion artifacts. Rather, due to the capacitor's complex physical geometry, materials and construction, it's more likely that the distortion produced will be an exceedingly complex amalgamation of both harmonic and a-harmonic signals of varying magnitude and phase. After all, unlike musical instruments, capacitors aren't intentionally designed to be "sound producers" that generate signals of any kind in the first place. Therefore, the odds are any artifacts produced by them aren't likely to resemble pleasant sounding musical signals either.
The MAGNETostrictiVE Effect:
The very physical shape of a FERROMAGNETIC material (such as are used in the cores of transformers and coils) can be changed simply due to a change in the state of its latent (at rest) magnetization. In the case of these circuitry components, a change in the state of their latent magnetization will occur due to the magnetic field developed from an applied electrical current passing through it. The force driving such changes in physical shape is called "Magnetostriction," and similar to the Piezoelectric Effect of Electrostriction , the reverse-magnetostriction effect is called the "Villari Effect." That being the case, it is the other source of distortion we are concerned with here.
All Pieces & Parts Become Sound Making Transducers:
So why does this effect occur and where does the sound actually come from? Well, it is generated by the two physical processes listed in the links above, and the end result thereof is that now our capacitor or transformer is also functioning as an electromechanical sound transducer, similar to that of a loudspeaker (albeit, a very poor one).
Even though in most low voltage/low current cases you cannot hear them, the same type of vibrations are being produced to some small degree by every metallic material that conducts electricity such as wires and connectors, or those that are exposed to it such as the insulation covering wires or the various insulating materials used to make electronic component parts. In fact, similar to the capacitor example above, the very silicon crystals that solid-state semiconductors are made from and many other related crystalline materials can exhibit a rather pronounced Piezoelectric Effect, which is often put to good use in the manufacture of small transducers such as "buzzers" and even "tweeters" that are sometimes found in certain loudspeaker systems.
All Pieces & Parts Become electrical Signal Generators:
For our purposes though, worse than the mechanical vibrations discussed are the consequences of both the Piezoelectric Effect and its counterpart, the "Villari Effect. In the transformer example, the Villari Effect is caused by the transformer materials experiencing the very same internal mechanical strain forces that cause the mechanical vibrations to be produced in the first place - magnetically induced physical displacement of the material. Alternating strain forces arising from the magnetic field produced by a flowing current cause the materials to become minute electrical generators of new, secondary signals that then get injected back into the circuitry of the device or audio component that is constructed from them. In our case regarding audio components, regardless of whether they are caused by Piezoelectric or Villari forces, these newly injected signals are directly related to the audio signals that gave rise to them, and in engineering terms such direct signal/artifact relationships are the standard definition of "distortion."
Therefore, if methods could be developed wherein these distortion artifacts could be measured separately, due to their extremely complex nature they would likely appear as a form of random noise, but in reality they are not. Rather, they are a REAL form of distortion that results from a real "CAUSAL" physical process of nature and are not the byproduct of any psycho-acoustic effects arising in the mind of the listener and/or "voodoo science," etc., that those of orthodox engineering persuasion might suggest.
Actually, the same phenomena is occurring within all of the wires, contacts and every other conducting material that is connected to the transformer in our previous example, but in most cases the effect is so small that we can't hear or see anything abnormal happening so we never become aware of it. Although, in cases where enough current happens to be flowing or the voltage is high enough, you might just observe the same effects elsewhere. Have you ever heard a circuit-breaker box in your home "buzz" when maybe you have a bunch of electrical appliances along with a whole-house air conditioning unit operating at the same time (high current)? Or maybe you have heard high voltage power lines up on telephone poles making a type of "buzzing" sound on a quiet, humid night (high voltage)? If so, those sounds are caused by the exact same physical principles we are outlining here.
So yes, strangely enough our transformer is now producing both AC voltages and currents that, if measured, one would discover are developed across (voltage) and flowing through (current) the wires connected to both its input (primary) and output (secondary) wiring. In addition, those same signals are also being transported to whatever else the wires happen to be connected to as well. Now you know why electrical engineers and technicians often refer to what they call a "dirty" AC power line, and why both audiophiles and audio professionals alike often spend large sums on AC power conditioners for their systems.
The fact is, every transformer and electric motor connected to the power grid that your home is also connected to are all sending similar "dirty" AC signals into your home's AC power system. Actually, the dirty power that comes into your home is mostly the result of different, more commonly recognized processes (like fly-back voltage spikes from motors starting up, etc.), but the above Electrostrictive/Magnetostrictive processes are still occurring as well - albeit at extremely low levels - right along with the other, more conventional noise sources.
The above then brings us to the point we are trying to make, and that is no matter how small they may happen to be, EVERY wire, circuit board trace, transistor, chip, capacitor, inductor, resistor, transformer... you name it... that has either a voltage developed across it (Electrostriction) or a current flowing through it (Magnetostriction) will physically move and/or vibrate to some small degree in response to all AC signal fluctuations it is exposed to. Of course, a DC voltage or current doesn't fluctuate, so those don't induce the effect. Conversely, AUDIO signals by their very nature are AC and do fluctuate, and it is in response to those fluctuations that the circuit components move about their rest positions ever so slightly. The effect becomes cumulative too, as each small part adds its own distortion artifacts BACK INTO the "whole," becoming part of we ultimately hear at the audio component's output.
In review of the above discussion, we find that whenever a given electronic part is forced to move physically in any way due to application of an electrical signal or otherwise, following Newton's third law of motion the effect reverses and now that same mechanically vibrating part becomes a small signal generator in return.
In fact, you can easily test the concept for yourself. Simply turn your amplifier on with the speakers connected and have a friend assist you. Place one of the speakers far enough away from the amplifier so that you can more easily hear any sounds that might emanate from it apart from any background noises. OK, the background noise we are most concerned about in this case is the sound of your friend lightly tapping on the metal cover of your amplifier, which he should do while at some distance away you place one ear very close to the speaker. Your friend may have to tap a little harder to generate the effect, but at some point he will be taping hard enough that you can hear a similar tapping sound coming directly from the speaker - as opposed to the sound traveling through the air and across the room to your ear.
That effect is direct evidence of Electrostrictive and Magnetostrictive forces we are discussing at work within the very circuitry of your amplifier and other components. Although the effects of these forces are extremely small when compared to the levels produced in the above experiment, they are still present to some degree at the output of virtually any audio component. That said, they are adding their own contribution and reduce the potential performance of the affected component to at least some small degree.
In light of all the above it should be clear by now that not only is every electrical part, wire and contact moving in response to the applied alternating voltages and currents they are exposed to within a given audio component, but simultaneously those same parts & pieces become small "signal generators" in their own right and thereby inject new signals into the circuitry that were never there before, and consequently... were never a part of the original source recording. Nevertheless, at the same time they ARE related in some very complex way (both harmonically and as sum/difference inter-modulation products) to the desired audio signal that initiated the whole process. Therefore, these artifacts will sound vaguely similar to the audio signal that gives rise to them, just like the tapping sound coming from the speaker resembles the physical sound of your friend's hand tapping on the cover of your amplifier in the experiment suggested above,
Nevertheless, taking into consideration so many parts, wires. connectors and even circuit board traces comprising the construction of even a modest audio device, one can well imagine that the spectral content constituting the cumulative sum of all these little "distortion generators" would be nigh unto infinite complexity.
To further complicate the matter, all of these "piece & parts" are mechanically connected together in some fashion, with the most obvious situation being the circuit boards the majority are soldered into (or "on," in the case of surface-mount construction). Due to the generally small physical dimensions (L. W. H.) and mass of these circuit components, the spectral makeup of the distortion artifacts that they generate will necessarily be limited mostly to those of the higher frequency ranges, with much (if not most) thereof residing out well beyond the limits of audibility.
OK, if they are physically vibrating, then those vibrations will travel through and within the circuit board material they are bonded to, and as a result spread out to all of the other associated circuit board components. One can easily envision and infinite variety of "resonances" taking place where parts of similar vibrational frequency vibrate in a synchronous fashion, with each reinforcing and amplifying the others. Adding to this, the circuit board itself will have an entire series of complex vibrational modes as well.
Now, one might argue that due to their relatively small physical dimensions (as stated above) combined with the relative high rigidity of the circuit boards commonly involved, it is likely that most of these resonant modes will reside at frequencies well beyond the limits of human hearing and therefore be be inaudible. So... what;s the problem? Well, some portion will always reside in the upper audio band to begin with, but those that are above can still impart a significant influence on sound quality. This is because of the well-known process of "Inter-modulation," wherein due to certain non-linear mechanisms within a system, two separate frequencies can "beat" together forming new "sum & difference" signals.
Specifically, when two different frequencies are added together, one Inter-modulation product that arises will be the sum of the two "parent" frequencies. In the case of the process we are concerned with here, the summed products will be even higher in frequency and therefore of lesser concern. On the other hand, this is not the case with the difference products. When one lower frequency is subtracted from a higher frequency, the Inter-modulation product will be the difference between the two parent frequencies and therefore much lower in frequency than either. The closer the two parents are in frequency, the lower the frequency of the inter-modulation difference product.
EXAMPLE: Say... an induced 49KHz resonant mechanical tone is beating with another at 50KHz. The sum and difference products will then be 99KHz and 1KHz respectively. While humans cannot hear a 99KHz tone whatsoever, we surely can hear one at 1KHz. So in this example our audio component will now have a small, residual output at 1KHz - coming and going over time down at the lowest levels of its dynamic range and likely working to mask the low-level details in the music.
Now just imagine: There could be many 10s, 100s, or even 1,000s of different resonant modes occurring all at the same time within a single given circuit board alone. In addition, not only can they be beating with each other, but also with the very audio signals that we are trying to reproduce accurately. The net profile of the entire process would require an advanced computer system and very complex FEM + SPICE "hybrid" modelling software to even attempt computer simulation.
As a bit of a side note, just imagine all of the little wire filaments within an audio vacuum tube vibrating away and transferring their vibrational energy to each other through their mechanical connections within the tube. Now you know why tube-based audio components have that "certain sound" so many enthusiasts are fond of. Thankfully, the overall effect is quite small though, or high fidelity sound as we have come to know and love wouldn't even exists.
As one can well imagine then, the combined energy of all these individual part vibrations, resonant modes and their Inter-modulation artifacts would sound grossly distorted with respect to the original audio signal that gave rise to them. Furthermore, these artifacts will also be directly aligned with their "parent" signal from a temporal standpoint; meaning they will only exist directly along side, and in a sense, ride "piggy-back" in time with respect to the parent - with the two essentially blending together.as one.
Residing only at the lowest levels of the dynamic range, in and amongst the true random noise that is also always present in every audio component, these extremely subtle distortion artifacts are metaphorically similar to miniscule invading viruses hiding among massively larger tissue cells. Similarly, this type of distortion is essentially masked by the much larger parent signal that induced it, and therefore cannot be easily detected by anything other than an ADVANCED INTELLIGENT FILTERING system. In other words... HUMAN HEARING. Without a doubt it is certain that they have been hiding there all along in our beloved audio signals with hardly anyone ever noticing, except that per-chance they should somehow suddenly be removed from the scene. THEN *maybe* we would notice their abrupt absence? Whaddaya wanna bet?
Logic dictates that the audible presence of these distortion artifacts would be to impart a "blurring" effect upon the parent signal so as to alter the makeup of its low-level harmonic structure. Well, those very fine details are what helps us to discern the unique timbre of individual instruments and voices. Other low-level information comprising the "ambiance" and imaging qualities of a recording can also be "masked" or blurred and/or obscured to a significant degree by these distortion artifacts as too, with a commensurate level of degradation thereof.
As we have discovered, the entire effect that these artifacts impart on the sound of a given component and/or system is so subtle that their very existence really only becomes apparent AFTER they have been REMOVED. The term of a "veil having been lifted" from covering up the micro-details in the music is commonly used to describe the resulting sound... and the effect is REAL - BASED IN SOLID SCIENCE AND NOT PSYCHO-ACOUSTICS!!! These artifacts DO exist and are part of our physical, causal reality as are all other conventional forms of distortion. Mystery solved - Academia be damned.
In summery, the sonic artifacts in question are a genuine form of DISTORTION because they are stimulated into existence by and are directly related to the original audio input signal that initiated them... just as all traditional forms of distortion. Although, the mechanism of their generation is far different from and infinitely more complex than those traditional field of Electrical Engineering typically concerns itself with. Therefore, except in the worst cases thereof, from the very earliest days of audio electronics manufacturing, as a wholethe entire realm of their activity and range of effect has been essentially ignored ... until NOW.
Fortunately, as a result of TDSS and our ongoing quest for ultimate audio performance, this is no longer the case. As a fundamental part of our D.R.R.T. service, through the application of various techniques & treatments we have achieved significant reductions in the levels of these undesirable distortion artifacts. When combined with the other treatments included, D.R.R.T. offers the devout audio enthusiast an unprecedented level of improved audio performance and value for their high-resolution audio system enjoyment.