Barnes and Noble Book Offer
 

 

  Background
 
Powrtron

 

In the early days of sound reproduction there was no sound amplification in a system. A needle attached to a stiff diaphragm caused the diaphragm to vibrate when the needle was on a moving sound track of a record. This provided a limited response with very low output. When records changed from cylinders to flat disks of shellac, RCA and others added an exponential horn to the diaphragm, mounting the diaphragm in the throat of a horn. This provided the first amplification of recorded sound (exponential horns can amplify sound under the right conditions).

  When receiving tubes were invented by Lee DeForest (DeForest inserted a grid between the cathode and plate of a diode), the resulting electrical power amplification made possible by vacuum tubes allowed the use of moving coil systems, attached to a cone. The speaker cone in these electrical systems moved enough air to eliminate the use of a horn (for most applications). This method of reproducing sound is the dominant method in sound reproduction even today. Minor changes such as separating the sound spectrum into two or more ranges (woofers and tweeters) have taken place, but in general the moving coil structure has persisted as the main vehicle for sound generation in home reproductive systems.
 
Inductance and capacitance
  In engineering terms, a loudspeaker is a transducer. It transforms electrical energy into mechanical energy and thence into acoustic energy (sound) by way of a vibrating diaphragm. As can be seen, this is transduction through three different forms of energy. Because of the limitations of the loudspeaker, energy is distorted as it passes from one mode of energy to another. Changing energy forms twice causes distortion to build from one mode to the other. Distortion on distortion as it were. The problem of the sound engineer is to reduce the distortion in each phase of transduction as much as possible. Gobs of money and time have been spent on this problem, seeking to reduction distortion of the system.
  Harry Olson and others have deciphered the electrical mesh created by a loudspeaker in its operation. As can be seen, this mesh contains a number of capacitances and inductances that reflect in the operation of the loudspeaker. These capacitances and inductances are REACTIVE power. Reactive power is "imaginary" in engineering terms. It generates no sound as such, but causes phase shift in portions of the radiated frequencies. If one examines a frequency response curve (constant voltage), one sees various sharp spikes and valleys from one frequency to the next. These peaks and valleys indicate phase shift. They also indicate the action of the mesh on the performance of the speaker.
  One human ear cannot detect stereo. It takes two ears. When sound strikes the ears of a listener, a phase shift will be detected and analyzed by the brain, making it possible to locate the direction from which the sound is coming. In other words, a person detects the three D location of a sound by phase shift. If a speaker reproductive system creates phase shift, the stereo effect is degraded. This is why most people have never heard stereo reproduction. Their sound system is too distorted phasewise to make it possible to locate the direction of a sound..
  Some of the peaks and valleys of loudspeaker response is caused by mechanical failures of the loudspeaker cone. These are generally referred to as "breakup modes" in the cone. Because few have loudspeakers without breakup modes, the manufacturers of speaker cones with breakup modes, tell the public that breakup modes are good for you. The fact is that breakup in your loudspeakers is as good for you as cigarettes. Only a very few speaker manufacturers produce speakers with no breakup modes in the used range, the SHOTGLASS speaker being one such speaker without breakup modes in the used range. Breakup modes create harmonic distortion, intermodulation distortion and phase distortion. It is the phase shift that destroys the stereo effect. Distortion of the wsve front is not the same as distortion of the stereo effect.
  There is another source of distortion that no one reveals with curves. This is power distortion. When a note is struck on a piano, a string on a guitar, a drum stick on a drumhead etc. a transient sound is created, the sound level goes from quiet to full power almost instantaneously. On a graph, the power curve goes from zero to maximum energy level almost instantaneously. This is not quite true, but for the purpose of description of the physics, this is what happens. Some refer to this phenomena as "transient" response. The initial surge in power has no sense of frequency, it is a wave front. Only after the vibration settles down, is there a sense of a frequency. This can easily be shown with a storage oscilloscope.
  Power response
  For various reasons, not much has been written about power response. The fact that various acoustic elements respond differently at different power levels has been recorded, but this is not the whole story. Steady state power response is not the same as sudden changes in power level (generally referred to as transient response). The response of various acoustic elements to different power levels, even such elements as microphones and amplifiers change their response as the power level is changed. The frequency response of various acoustic elements under transient conditions, is seemingly neither stated by most nor understood.
  If an acoustic system (say pulse generator, amplifier, loudspeaker and microphone) is placed in an acoustic sound chamber (anechoic) and observations are made (outside the chamber) on the performance of the system, (using various lengths of pulse chains), an unexpected result is obtained. The minimum length of pulse chain that can produce a recognizable sound is four units. A pulse chain of less than four units sounds like a thud, not a sound. Mathematicians will tell you that a frequency can be determined with only a portion of a wavelength. This may work mathematically; it does not work physically, showing once again the limitations of mathematics.
 

What does this mean to a listener? It means that a transient, like a drumstick, a piano note or the pluck of a guitar string is not easy to distinguish, either by a sound system or a listener. The sudden change in sound level is observed, but unless the transient persists for more than four wavelengths, no note will be recognized. The power envelop of a transient is difficult to analyze. One of the differences between tube amplifiers and transistor amplifiers is how they handle transients. An ordinary tube amplifier rounds the wavefront powerwise (the response is not instantaneous), while a transistor amplifier responds instantaneously (voltagewise). Seemingly, the instantaneous response as generated by a transistor amplifier is not as pleasing as the rounded response of a tube amplifier.
This is one reason tube amplifiers are making a comeback.

  Not all amplifiers are equal
  As is well known, when an electrical transient impinges on a mesh, it causes ringing in the mesh. This ringing is not related to the frequency of an impressed sound, but to a system resonance response of the amplifier-loudspeaker system. A transistor amplifier (or any amplifier with a voltage feedback loop) will suppress the voltage elements of the ringing. What doesn't happen is a repression of the internal currents in the loudspeaker mesh. Ordinary amplifiers do not respond to current as such. The circulating currents in the loudspeaker mesh generate magnetic fields in the loudspeaker voice coil. These unwanted voice coil currents create distortion. Because transistor amplifiers have gobs of voltage feedback (many times more than tube amplifiers), the seeming result is more distortion in the ringing loudspeaker mesh than that created by tube amplifiers. But this is not the whole story.
  As stated above, a loudspeaker, electrically has a mesh circuit, containing inductances and capacitances s well as a resistive element. Part of the impedance of the mesh is motional impedance, caused by the motion of the speaker voice coil in the magnetic gap of the speaker. The motional impedance has a mechanical and an electrical aspect, but the electrical aspect is the part of the impedance that affects the amplifier.
  A loudspeaker is a power device. It takes input electrical power and converts it to output acoustical power. The motor of most loudspeakers is a vibratory wire coil imbedded in a strong magnetic field. As the magnetic field of the coil varies, it causes a magnetic reaction that results in a one degree vibratory motion. As the electric coil (called a voice coil) is attached to a suspended diaphragm (called the speaker cone), the motion of the voice coil is transmitted to the speaker cone, causing a change in air pressure that the listener identifies as sound. Current audio amplifiers control the voltage out of the amplifier (voltage feedback), trusting to luck that the speaker responds properly. (A simple voltage feedback loop does not sense the motion of the speaker voice coil. That this system works at all is a credit to the ability of the human ear to identify a sound pattern, not in the ability of the system to recreate the original sound pattern faithfully. Proof of this---have you ever heard a recording of a loudspeaker output? Sounds like an old Edison phonograph.
  A simple electrical law states that the current in a loop is constant throughout the loop. In the operation of a loudspeaker, the voltage across the speaker terminals times the in phase current is real electrical power input and will be reflected in the output of the speaker. A voice coil has inductance. This inductance is a phase lagging electrical parameter. What it means is that when a voltage appears across the speaker terminals, there is a time lag (caused by the voice coil inductance) before real acoustic power appears. This means is that impressed transients suffer a time lag before acoustic power appears. Heavy cones create mechanical inductance and voice coils create electrical inductance. Both factors can inhibit transient response. This time lag is sort of an inertia. In reality, real sounds do not suffer from this inertia.
  The POWRTRON "FF" model has made possible a partial cure for this inductance (inertia) problem. It is called field forcing. The technique is common in the electrical power field. A way has been worked out to apply this industrial echnique to loudspeakers. The FF amplifier nullifies the (inertial) effect of time lag caused by heavy diaphragms and voice coil inductance. There is a noticeable improvement in transient response when the new FF amplifier is used in conjunction with loudspeakers. In a real sense, the FF is not an amplifier but a servo mechanism. The FF senses the inertia and corrects for it. The improvement is noticeable no matter what kind of speaker is used. Of course, speakers like the Shotglass (with no modal breakup) will sound better than ordinary speakers, but that is to be expected.
  © Copyright 2000 -2004 Stanley F. White