Jul 18 2011

Pique the Geek 20110717: Loudspeakers

In electronics, a loudspeaker is what most people just call a speaker, the device that converts electrical signals to sound.  They can range from very simple to very complex designs, with variations in cost from just a few cents to thousands of dollars.

All practical loudspeakers are electromechanical devices, using an analogue electrical signal to make the loudspeaker components to move in such a manner as to in turn move air (usually, although other media can be used for purposes other than human perception) and thus make a sound.  For human hearing, air is almost always the medium used.

Loudspeakers are one of the few modern electronic devices that are analogue only.  In other words, a truly digital loudspeaker does not exist except in a few research laboratories and they are not very good.  It is interesting to me that the final stage of reproducing sound is firmly rooted in the 19th century insofar as basic technology in concerned.  This discussion is limited to electromechanical loudspeakers.  Purely mechanical ones are much older than electromechanical ones.

There are many different basic concepts and many more different designs for loudspeakers, but two items are common to all of them.  The first is some sort of transducer that converts the analogue electrical signals into motion in a relatively small, usually high mass assembly.  This is coupled to some sort of larger, lower mass material that can move a relatively large amount of air, called the driver.  I am not really happy with those terms, because they are a bit nebulous.  If I were to give them names, I would call the transducer the driver and the air moving part the resonator, but that is just my personal preference.  I shall use the accepted naming conventions for the rest of this piece.

Before we get into the Geeky parts, think about how many loudspeakers you have in your everyday life.  Let us start with things that you are likely to wear.  Electronic watches often have loudspeakers for alarm setting.  All kinds of telephones have them, and many of us wear a cellular telephone on a belt.  Radios, TeeVees, and stereo systems HAVE to have them, and my microwave oven and kitchen range also have them.  Electric alarm clocks, computers, and many, many other items have loudspeakers associated with them.  As a matter of fact, almost anything that you have that makes sound has a loudspeaker.  Some of the few things that intentionally make sounds NOT associated with loudspeakers include door chimes, mechanical bells (like on old telephones), and buzzers come to mind.  Car horns are very much like loudspeakers, but do not QUITE qualify.

The very first loudspeaker to my knowledge was developed by the brilliant, and self taught, German inventor Johann Philipp Reis.  He devised a number of models that used the concept of a platinum spring that would “jump” to a thin platinum diaphragm as the electrical contacts were made and broken.  It was  not a very good design, but worked well enough for pure tones and simple music.  Voice is much more difficult, because of all of the discontinuities in it.  He was trying to invent a telephone, and actually did, using the German term telephon, but his loudspeaker just did not reproduce voice well enough to be practical.  However, the concept DID work, but just not for voice.

Graham Bell produced the first loudspeaker that could faithfully (for the time) reproduce human speech, so he really invented the loudspeaker that really worked.  That was the contribution to the modern telephone that he really made.  The concepts had already been proven, but the technology was not up to the task.  Bell was the pioneer of the attached membrane to the driver (once again my terms), that made the loudspeaker possible.

There is much controversy about who invented the telephone and when.  But this is not really about the telephone.  I use these historical facts to illustrate that the telephone was the first driving force to perfect the loudspeaker.

Most loudspeakers use the moving coil design, meaning that a coil of wire is attached to the sound radiating device, and the coil is suspended in a constant magnetic field.  Current carrying the information for the waveforms to be reproduced causes a varying magnetic field in the coil, which makes it move because of interactions with the static magnetic field.  This movement produces the sound.  Most modern moving coil loudspeakers use permanent magnets for the static field, but this was not always so.  Back in the days when steel was the best magnetic material to be had, electromagnets were used to produce the static field because steel magnets were just too weak.

Most moving coil loudspeakers use ceramic magnets due to their relative high performance and low cost, but other magnetic materials have been and are being used, such as Alnico magnets and the new rare earth magnets.  These have better performance, but are much more expensive than ceramic magnets.

A very important design element in moving coil loudspeakers has to do with the frequencies of sound to be reproduced.  The human ear has its maximum sensitivity at around 5000 Hz, so most general purpose loudspeakers are designed to reproduce frequencies around that one, dropping off rapidly within a couple of thousand Hz on either side.  These are fine for telephones, most TeeVee applications, and other applications where the voice is the most important information to be carried.  Let us consider the physics behind this for a moment.

Let us consider first a typical home stereo system consisting of three loudspeakers for three different frequency ranges.  Many systems break this down further into more ranges, but the fundamental are similar.

Very low frequencies, to be heard, require the movement of LOTS of air because our perception is not very acute in this range.  Most people can hear down to around 20 Hz but honestly it is more “feeling” than it is hearing.  At those very low frequencies we, if the volume is high enough, actually sense the pressure and rarification waves by touch rather than by hearing them, and as frequencies drop, the touch is by far more important than hearing.  Thus, to reproduce these low frequencies requires a rather large loudspeaker so that enough air can be moved to be sensed.  This requires a robust design and materials set, because of the shaking involved.  However, since these frequencies involve relatively slow motion in the driver, the components can be rather sturdy.  Remember, heavy things accelerate more slowly than light things, so it is easier to design a low frequency loudspeaker with heavy components than it is to use heavy components for higher frequencies.  For home stereo, these loudspeakers are usually called woofers down to 200 Hz and subwoofers for lower frequencies, although those limits are rather arbitrary.

For higher frequencies, up to around 5000 Hz, a mid-range loudspeaker is used.  They are essentially useless for lower frequencies, but are optimized for this range.  These are generally smaller than woofers and are lighter, because the motion of the driver is much faster than that for a woofer.  The same principles hold, just the geometry being optimized for the frequency range being optimized.  When I talk about it being light, I mean only the moving parts.  The magnet has no restriction for weight, since it does not move.

For frequencies heading towards 20,000 Hz (pretty much the upper limit of human perception) even smaller, lighter loudspeakers, known as tweeters, are used.  They are difficult to design because of the extreme rapidity of motion of the driver, so they have to be both light and robust.  The magnet is generally larger in proportion to other speakers, because the sound radiating components have to be pulled back to “neutral” much more quickly than for lower frequency applications.

Now, with three frequency ranges for the loudspeakers, some sort of mechanism has to be devised to direct the proper frequency range to the appropriate loudspeaker.  This is done with an electronic filter that removes, say, low frequency signals for the information to be reproduced by the tweeter.  This helps to eliminate signals that the loudspeaker can not reproduce due to its design and that would cause distortion.  It also improves overall volume because the full signal is directed to the proper loudspeaker rather than be divided betwixt all three.  This is called a crossover network, and each manufacturer uses different specifications depending on their particular loudspeaker design.

There are also what are called full range loudspeakers, but are NOT full range at all.  They are compromises to do the best job possible for reproducing sound over a fairly wide frequency range, but are hardly what one would call high fidelity.  As mentioned earlier, they are fine for AM radio, lots of TeeVee, bullhorns, and the like.  They are generally inexpensive and do not have very good sound, although higher quality ones can sound better than cheap three loudspeaker arrangements.  They require no crossover network, because manufacturers just put filters in the amplifier driving them so that frequencies beyond their range are simply not supplied to them.

As I said earlier, the vast majority of loudspeakers in use today are the moving coil type, also called electrodynamic loudspeakers, although there are a few other types of note that we shall discuss very briefly later.  Now, we have a magnet and a voice coil, with the signal passing into the voice coil, causing it to vibrate.  However, it has to be attached to something to move enough air to make an audible sound.  Most home loudspeakers are of the cone variety, meaning that the voice coil is attached (usually by strong adhesives) to a cone of some composite material, and this cone actually moves the air.  There is a variation of this called a dome driver, but it basically is the same thing with a different geometry.

The cone (or dome) designers must attempt satisfy three mutually exclusive requirements simultaneously.  First, it has to be as light as possible so that it can respond rapidly to changing signals.  Second, it has to be as rigid as possible so that it does not “flop” around, causing distortion.  Finally, has to be well-damped, meaning that it rapidly stops moving when signals stop to prevent distortion.  No single material can meet all of these requirements, so attempts are made to provide the best overall performance consistent with cost.  It turns out that stiffness is inversely related to damping, so extremely still materials are not suitable, so you do not see metal cones.  Most cones are primarily made of paper, which is good for lightness and damping, but poor for rigidity.  Several things can be done to improve stiffness without too much sacrifice in the other two parameters, such as using composites added to the paper, special coatings to modify stiffness, and other tricks, some of them proprietary.  By the way, practically all “paper” cones and domes are treated with a coating to make them immune to atmospheric moisture.  Paper is quite hygroscopic, so the mass (and the stiffness) or a cone or dome would change with different humidity levels without such a treatment, thus causing degradation in performance.

I have not yet mentioned the chassis into which a loudspeaker sits.  There are several requirements, the most important being that it is extremely rigid, to keep everything precisely aligned.  Another important property is for the natural resonance frequency of the support to be far from the frequencies that the loudspeaker reproduces so that it does not vibrate and so introduce distortion.

The next most common type of loudspeaker for home use is the horn driver.  In many respects they are similar to the cone design, but the cone is typically much smaller for a given volume because of the phenomenon called acoustic coupling, then horn being the coupler betwixt the small cone and the air.  The bottom line is that you can do a much better job of optimizing the relatively small cone for the frequency range desired, and by coupling this cone to a properly designed horn, get much more efficiency (volume of sound per watt of power) than with just a larger cone.  I have a pair of Klipsch Heresy loudspeakers that I bought in 1978 that I can drive to the point of running me out of the room, at only about half power (100 W) from my beloved Kenwood KA-8100 direct coupled integrated amplifier (I call it “Old Sparky”), yet they sound perfectly fine when driven by a standard car stereo (not a hyped up one), except for the bass.

Granted, these are high quality loudspeakers.  Here is a picture of one just like it from the Klipsch website.


The reason that they are called Heresy is that this was the first model of loudspeaker build by Klipsch that used a cone only driver (for the bass).  It is not possible to build a horn bass driver in such a small enclosure because of the size of a bass horn, so they compromised and used the cone for the bass.  As I said, I bought them in 1978 for $243 apiece (big money at the time).  They are now discontinued, but the nearest derivative of them is still built in Hope, Arkansas by Klipsch for special order (eight week wait) for $799 each.  These babies are HEAVY, too!  Lots of magnet and cast aluminum chassis mass, and the heavy plywood enclosure adds a lot as well.  Paul Klipsch was an acoustical engineering genius, and I should do a piece on him for Popular Culture sometime soon.

I promised to mention a couple of other technologies other than electrodynamic loudspeakers.  For high fidelity applications, probably the best is the electrostatic loudspeaker.  It uses a completely different driver design, not based on a moving voice coil.  In this design, a thin polymer membrane is surrounded by a very high voltage electric field, and the signal is passed through the stators that also supply the field.  The membrane vibrates due to electrostatic attraction and repulsion.  They are very light and thin, and offer excellent reproduction since the membrane itself is the radiator, and sound great.  One disadvantage is that whereas my Heresies sound good even at five watts, forget about electrostatics unless you can push way over 100 watts.  They also attract dust because of the high voltage field, and sometimes arc and spark.  They would be a good combination for Old Sparky!  They are also physically large in two dimensions, ofter several feet square.  Whilst they are great for midrange and high frequencies, they are not good with bass, so usually an auxiliary woofer is added.

Although there are lots of other kinds of loudspeakers, the last one that I shall mention is the piezoelectric kind, because they are very common.  In this design, a piezoelectric material (usually some sort of ceramic) is directly connected to the signal.  A piezoelectric material has the property of generating an electrical output when distorted (the “electronic” disposable lighters use this property:  a small spring loaded hammer hits the ceramic, and a spark is emitted, a direct conversion of mechanical energy to electrical energy). Conversely, when an electrical signal is applied to such a material, it changes shape, and changing shape is producing sound.  They are commonly used on digital watches for alarms, and lots of inboard computer speakers are this kind.  The are pretty efficient, but do not provide the fidelity that electrodymanic loudspeakers do.

Well, you have done it again!  You have wasted many einsteins of perfectly good photons reading this noisy piece!  And even though the Republicans realize that they are playing chicken with the worldwide economy when they read me say it, I always learn much more than I could possibly hope to teach by writing this series, so keep those comments, questions, corrections, and other feedback coming.  The comments are almost always the best part of these pieces.  Tips and recs are also welcome, of course.  I shall remain here as long as comments warrant this evening, and shall return for Review Time tomorrow after Keith’s show.  He STILL has not contacted me about that science adviser gig, but I am sure that he is busy getting his show off the ground.

Warmest regards,

Doc, aka Dr. David W. Smith

1 ping

Comments have been disabled.