Krkus: Very nice. I like the speaker damping - stuffing leading to damping factor for an amplifier etymology. I also like your reasoning based on the multiple tap amplifier technology of bygone times. |
Actually Musicnoise, I think the origin of the term "damping factor" for audio amplifiers has much humbler roots than classical Control Theory.
In these days of direct-coupled amps, it makes sense to specify output impedance, but when virtually all audio amplifiers had multiple output taps . . . I can see where it may have been handy to express this as a single, approximate number for all output impedance configurations.
And differences in the amplifier output impedance can logically be likened to differences in the amount of stuffing inside a speaker box . . . or the compliance in a speaker cone . . . and if we call this "speaker damping", then "damping factor" for an amplifier is actually pretty concise and specific, especially in the world of published consumer electronics specifications. Just a bit anachronistic today. |
For the accepted engineering definition of "damping factor" see for eg. Automatic Control Systems, Kuo; or, Digital Control System Analysis and Design, Phillips and Nagle. Both published by Prentice Hall. The first edition of Phillips and Nagle has a very nice graphical description of how the transient response of a system changes depending on pole location - of course in the z-plan rather than the s-plane seeing as in that instance the analysis is based on on a discrete rather than a continuous analysis (z in that analysis has nothing to do with impedance but comes from the name given for tranforming number sequences to the frequency domain).
Once you review the definition accepted by the engineering community, you will see that damping factor has to do with a system, not merely one component of the system or one set of measurements of a system. The point is that the term, as it has been 'borrowed' is a marketing tool. This is common in the marketing to audiophiles - see the recent thread on the guy who was trying to sell his product to correct for the "doppler effect". The doppler effect has a real scientific definition and by that definition it makes no sense to talk about the doppler effect occurring within an electronic amplifier.
The appropriation of terms in this manner works in marketing like this - from the naive purchasers viewpoint "Oh, the (name the borrowed term) is something I have not considered in my system. It certainly sounds official, and it's "science" so it must be important. Now there is something that I can purchase that will deal with this. If I don't correct for the (name the borrowed term) then my system will not be optimized. I better buy it. Then I will have dealt with another aspect of system degradation." Of course when engineers and scientists debunk the practice, the response shifts to "Oh, I don't know why it works it just does - you can hear it - although sciene and engineering may be able to send man into space they cannot explain the complex world of home audio" or "but these are not simple sine waves - music is complex." |
Wikipedia definition:
In audio system terminology the damping factor gives the ratio of the rated impedance of the loudspeaker to the source impedance. Only the resistive part of the loudspeaker impedance is used. The amplifier output impedance is also assumed to be totally resistive. |
My guess is that the term "damping factor" was at some time in the past appropriated by audio manufacturers for marketing purposes. The term is defined in automatic control system theory where as the product of the damping ratio and the natural undamped frequency. The damping factor is the distance in the left half of the s or plane along the real axis. It is a term specified wholly in terms of mathematics and is completely generic. It's application is in determining the transient response of a system to a step imput and requires that the system first be expressed in terms of a characteristic equation. Factors that are found in a typical electromechanical system include the torque constant, load intertia, amplifier gain, armature resistance, back emf constant etc, acted upon by appropriate factors.
In short, damping factor as it is understood by engineers in all disciplines is not reduced to a simple ratio of two numbers. Since the term has an actual scientifically accepted meaning, my guess is that the term was appropriated to lend scientific status to what someone was trying to sell.
That all being said, knowing the input and output impedances of the components of a home audio system is helpful in choosing which items to connect together in a system, how to connect the items, and how to avoid or address problems. |
Once above a certain minimal value, who cares?
Exactly. However one caveat is that you are using a speaker with a "reasonable" impedance curve. Some speakers designs dip down to 2 ohms (not a good thing)- in this case a much higher damping factor than 10 will certainly help. The Speaker has the last word here. I have heard a speaker alledged to have a 'Q' of .707 and it sounded almost bass thin. Yes that is how they will be perceived because many speakers are not designed this way. A Q of 0.707 is "critically damped". It means the woofer goes the most quickly to zero after power is removed without any overshoot (no added oscillations or extra bass notes). Some speakers are designed with a higher Q. This allows them to have much more bass response (higher efficiency with a typical hump in the bass response on a freq plot) but the signal continues to oscillate after the power is removed. (It also allows for a smaller box to achieve good bass output) This means transients and decays are not represented properly (timbre will be wrong) but you get a pleasing thick and impressively powerful bass sound (it sells in A/B shop floor scenarios). Sound is two things amplitude and duration - the longer the bass note lasts the louder or more prominent (impressive) it will seem in the mix. Does a low 'Q' speaker have what I have heard referred to as 'bloom'? A low Q speaker is over damped. It will be rather inefficient and will require lots of power to drive it. (This type design is extremely rare) The response will go to zero when power is removed and it will not overshoot, however, it will be sluggish compared to "critically damped". Think of a a typical North American storm door and how it closes very slowly - this is overdamped. It will sound even thinner than "critically damped" a very dry and tight punchy sound given the right copious amount of power to control it. It will not sound like "bloom". Bass "bloom" or one note bass would be from a Q of say 0.9 - 1.2 (actually so common that this may be percieved by many as being "correct" sounding bass whereas Q = 0.707 will be perceived as being bass light or wrong sounding bass) |
Shadorne, At least you get it.
Once above a certain minimal value, who cares?
The Speaker has the last word here. I have heard a speaker alledged to have a 'Q' of .707 and it sounded almost bass thin.
Does a low 'Q' speaker have what I have heard referred to as 'bloom'? |
Rleff - Keep in mind what "damping" is, as opposed to "damping factor". If you look at the response of a system, any system, to an input signal, the question is how well does the output follow that input - both in terms of time and magnitude. At the extremes systems that are not properly damped are will either be too slow or will have too much overshoot. Damping can be accomplished by mechanical, electrical, pneumatic, etc, methods. Damping factor on the other hand, at least in terms of audio amplifiers (as opposed to the definition found in control system theory) is simply a ratio of impedances and at least one of the impedences is never fixed.
Jj2468 - in response to the other half of your question - a low output impedance will be somewhere less than 0.04 ohms. Most solid state amps will be there without a problem. You are more likely to have high output impedances with tube amps. |
Damping is very important but it is more related to speaker design than anything else. It is determined by Q factor. Above 10 you start to get diminshing returns on an amp - above 50 you are more less getting close to inaudible improvements in damping from going higher. A speaker will have a Q of .707 to be critically damped. Many speakers have a bump in the bass and are underdamped - no amplifier can compensate for the sloppy resonant sound of this type speaker. |
Jj2468 & Rleff : In general, the specs for output Z will be listed in the owners manual of the amplifier or on the manufacturers web site. One should also note the input Z as this can come into play when interfacing a preamp or source to the amplifier.
As to damping factor, a damping factor of 200 or 2000 will be indistinguishable in practice. The damping factor is just Zspeaker / Zout amp. One could say that this is equivalent to specifying Zout of the amplifier, and, technically that is true. The problem that I have with paying a lot of attention to the damping factor is that it is a step removed from the fundamental parameter, the output Z of the amp. One must keep in mind a caveat - the damping factor has to be stated with regard to a given load impedance, usually this is 8 ohms. But, no speaker has a constant impedance from 20Hz to 20 kHz. In deciding whether the changes in speaker impedance over frequency will affect the sound, it is simpler and more straightforward to look at speaker impedance curve or the minimum and maximum speaker impedance specs and then compare those to the amplifier output Z. So long as the ratio is better than 100 or so, you should be fine. Whether it is 200 or 2000 is not going to matter. When it comes to figuring out how to select components for a system and what could cause problems in a system, thinking in terms of base parameters rather than derived parameters makes life easier because it lessens the tendency to unnecessarily complicate the the issue. |
'Z' means impedance in 'electronic speak'
Amps damping factor is another one of those minimal specs.
Above a certain minimal value they don't mean much except to the advertising copywriting guy.
Besides, the amp doesn't 'absorb' or 'Damp' or 'control' the Back EMF from the woofer. Rather, the woofer controls itself.
You can test this idea by thumping the edge of a woofer with a finger. Than, short out the leads to the driver with a piece of wire. Thump again and the sound changes lots. this is the woofer damping itself. Low input impedence of an amps outputs just allow the current to pass thru, back to the woofer.
More (arguable) is the speaker systems design 'Q', which also is an expression of damping. What is called 'critically damped' design of Q=.7 will give nice tight bass even with a minimal damping factor amp. A speaker with a hi 'Q', say 1.2 or so, will sound sloppy, even with a high damping factor amp. |
Musicnoise- I was always under the impression that damping factors are important for controling the woofer and being able produce tighter bass;would that be correct? |
"Find an amplifier that has low output Z..."
What is "low output Z" and how do you find it? Jeff |
Dazzdax: For the most part the slew rate of just about any amplifier you are using at home is going to be a non-issue. Any signal, no matter what it looks like, can be broken down into a series of sine waves at different amplitudes and phases (Fourier analysis is the mathematical method for proving this). You can only hear the components out to 20 kHz. The max rate of change of a 20 kHz sine wave is 7.5 v/ usec
If you go through the calculations and then double the value for a safety factor, you arrive at an amplifier with a slew rate of 10 v/usec a capable of accurately reproduce a 20kHz sine wave while providing 400 W rms into 8 ohms. So whether you have an amplifier with a slew rate of 10 v/usec or 10000 v/usec, makes no difference.
If you want to get extreme, an amplifier with a 80 kHz bandwidth and providing 1000 W rms for an 8 ohm load will accurately reproduce all of the frequencies in that bandwidth if it has a 64 V/usec slew rate.
In short, slew rate is not something to brag about in an amplifier. Unless the design is simply terrible, there is not going to be a difference in the sound - the amplifier will be fast enough. The placebo effect is of course another story. More important are such specs as output impedance, distortion, and current capability. Find an amplifier that has low output Z, a low distortion spec and a wattage rating for driving 8, 4, and 2 ohm loads that comes close to doubling for each halving of the load Z.
Pay attention to the specs that make a difference. High slew rates and high damping factors are meaningless. |
So rise time and slew rate are not related to each other? Does an amplifier that has very high slew rate very good transient response? Does it sound more dynamic?
Chris |
One more point as far as interrelationships. If it is not already obvious from the previous posts, if you are using an amplifier to process pulses you would first determine the rise time and fall time of the fastest changing pulse and the change in volts of that pulse, that you expect to encounter and then select your active devices based on that time. That is a starting point. In many instances the passive components that you select and the pcb layout and physical construction may also need to be considered, depending on those times. |
Rise time is defined as the time it takes a signal to go from 10 percent of it's peak value to 90 percent of its peak value. For example, if you have a 0 to 10 volt pulse, the rise time is the time it takes for the pulse to go from 1 volt to 9 volts. For a 0 top 5 volt pulse, rise time is the time it takes the pulse to go from 0.5 volts to 4.5 volts. Fall time is similiarly defined, but in the other direction. Rise and fall times are just that, time measurements specified by a single unit. Slew rate is a units per time (eg. volts per time) measurement, specified with two units; eg. volts / time.
Slew rate is typically used to describe an amplifier (look at a spec sheet for an op amp). Rise and fall times are typically used describe signals. |
Thanks for the links and explanation guys!
Chris |
Chris - slew rate is a slope of the signal going up or down expressed in V/us (rate of change). Rise time is only for known amount of volts going up (fall time for going down).
We use slew rate when voltage doesn't have any particular range of interest (output of an amp etc.) and we just want to know how fast is it changing. Another case might be when something depends, by definition, only on the rate of change - like transmission line effect.
Rise time and fall time are used mostly when voltage range is already defined like 5V logic or 3.3V logic etc.
Slew rate (as well as rise or fall time) can be applied to any analog quantity like current, resistance, light etc. |
(http://www.amplifier.cd/Tutorial/Slew_Rate/SlewRate.htm) (http://electronics.wikia.com/wiki/Amplifier#Bandwidth_and_rise_time) Click on, "bandwidth and risetime" second site. They are very similar, but not identical. These two threads should give you some insights. |
