Posted by Al Sekela on February 06, 2003 at 08:56:07:
In Reply to: Thought policeman speaks posted by Lew on February 05, 2003 at 14:05:52:
Lew,
Thank you for your continued questions. I'm not being clear on this topic, and I want to make sure all interested parties understand my ideas. Please keep asking until I make it clear.
"Damping" is the action of dissipating energy. Any system that takes energy in one form and converts it to another in a partly reversible way may oscillate under certain conditions. This is true of chemical, electrical, mechanical, and even quantum-mechanical systems, but we are here interested in speakers. Damping in speakers is the combination of forces that act to stop the oscillation. These forces dissipate the energy stored in the speaker. Some of the damping may be built in to the speaker components or come from air resistance, but most of it is due to interaction with the amplifier.
The basic problem with speakers is that they are made of real materials, with finite mass, and their parts can oscillate. This is bad, because the oscillation produces sounds that were not represented in the driving signal. Damping is good if it minimizes oscillations without interfering with the speaker's primary job, which is to convert the driving signal into sound. Damping is bad if it interferes with the speaker's job more than it helps by reducing oscillations. There is some degree of damping that is a perfect balance of too little and too much. If you recall the Dan Tomcik article, this is referred to as "critical damping." It is perfect for woofers, because it brings the woofer cone to a stop in the absence of signal in the least possible time.
If this sounds too confusing, just think of what Americans call 'shock absorbers' on automobiles. When the shocks wear out, they lose their ability to dampen the motion of the wheels with respect to the car. The wheels bounce up and down so much that the car is difficult to control. This is called underdamping.
On the other hand, imagine what happens if someone installs shocks that are much too stiff for car. The wheels don't bounce with respect to the car, but the control still suffers because the car bounces (and the ride is terrible!). This is called overdamping.
For every car, there is a perfect degree of damping that gives the best control (this may not be the degree of damping chosen by the car designer, because most drivers are not racers and want more ride comfort). This would be the "critical" damping for the particular car suspension.
Dan Tomcik was concerned with bass response from cone woofers driven by magnetic interaction of voice coil currents with a fixed magnetic field. Such a motor also acts as a generator: when the signal goes to zero, any cone motion causes current to flow in the voice coil that, in turn, acts to stop the cone motion. This generated current provides the damping force. It comes from the energy stored in the mass and spring of the woofer cone and suspension. The amplifier acts to control the damping by permitting the current to flow to a greater or lesser degree. This property of the amplifier is called its "output impedance."
Because cone woofers are so common, it has become customary to refer to amplifier "damping factor" with respect to cone woofers. A "high" damping factor causes a larger damping force for a given cone woofer than a "low" damping factor. As Dan Tomcik observed, it is possible to construct a cone woofer with so much voice coil resistance that there is no way to achieve critical damping with any stable amplifier, but for most woofers, there is a perfect value of damping factor that achieves critical damping.
Tomcik developed amplifiers under the Electro-Voice label that had adjustable damping factor so that this value could be dialed in by the user. This was copied by other amplifier manufacturers in the late 1950s and early 1960s, but the feature died out when transistor amplifiers came along. It probably died out because most users did not have the equipment to determine exactly where the control should have been set to achieve critical damping. Tuning the bass response by ear is difficult, and calibrated microphones and oscilloscopes would have been very cumbersome and expensive accessories for the average audiophile in the vacuum tube era.
However, the concept of damping factor adjustment has survived in the Carver and Wolcott amplifiers. There are "current source" and "voltage source" outputs on some Carver (Sunfire) designs, where the "current source" is simply the voltage source in series with a resistor (this is what I understand, but I have not taken one of these amplifiers apart to confirm), and the Wolcott amplifier provides a switch which controls the feedback network to set the output impedance to higher or lower values. For the Wolcott amplifier, the "high damping factor" switch position causes the lower output impedance, and vice versa. This nomenclature is with respect to cone woofer damping as described above and in the Tomcik article.
Electrostatic speakers have very light membranes, and the energy stored in them is much lower than that stored in cone woofers. However, the membranes are not stiff, so they can vibrate with many modes. These modes are like the overtones of a string under tension, but because they exist in two dimensions rather than one, there is a much larger variety of shapes they can assume. The overtones of a string are limited to the whole-number harmonics of the half-wave fundamental tone, while the overtones of a panel can have a bewildering array of values, some of which are very "unmusical."
My understanding is that our perceptions of harmonic distortion are conditioned by how "musical" the distortion overtone relationship is to the driving signal. Panel resonances that do not mimic the natural overtones of a string, therefore, do not have to be at a very high absolute level to create an irritating effect to us. In this case, even a little damping is better than none, if it reduces our perception of these "unmusical" overtones.
Now, how do amplifiers work to dampen electrostatic speakers? I discussed this in an earlier posting. The driving force in an ES speaker is voltage, not current. Membrane oscillations will cause image charges to appear and disappear in time with the oscillations on the stators, and these image charges have to move as current through the amplifier to go back and forth between the stators. More amplifier resistance will act to dissipate more energy from these moving charges, while less amplifier resistance will let them flow with less dissipation, and zero amplifier resistance (which can be achieved with the use of feedback), will let them flow unimpeded.
In this case, the Wolcott amplifier with "high damping factor" selected will have the lowest output impedance and have the least damping effect on the ES speaker. Moving the switch to "low damping factor" will increase the amplifier's output impedance and allow it to provide some damping influence on the higher panel oscillation frequencies.
My speculation is that Atma-Sphere OTL amps and Wolcott amps with "low damping factor" selected sound better on ES speakers because they are able to dampen some of the more annoying high frequency panel oscillation modes. Transistor amps and Wolcott amps with "high damping factor" selected have less influence on the panel modes and let their harshness through.
It is not necessary to achieve "critical damping" of all possible panel modes. There are so many of them, and at so many different frequencies, that this is not practical. However, if the worst offenders (the ones that do not relate in a way we perceive as "musical") are dampened significantly, we will hear the speakers as more clear and life-like.
If anyone thinks Sound Lab speakers do not have panel vibration modes, please reread Dr. West's description of distributed resonance carefully on the Sound Lab website. Dr. West has used panel modes to control the speaker's impedance, but within each subpanel selected to distribute the resonance, there are higher-order modes. It is the damping of these modes that is described above.
-Al