Speaker Design
and DJ's Part II of III
by
Larry Mundy
April 2004
The following
article by Larry Mundy will appear in a forthcoming book on the
design and construction of pro-audio cabinets. Larry Mundy retains
ownership of this article and it is republished here with his
permission.
[Editor's Note:
At one time or another, "distortion," "clipping,"
and "blown speaker" are words you will hear in almost
every DJ booth. A fundamental understanding of speaker design
may improve the quality of your sound and ability to move a crowd,
while avoiding system problems. This month's tip is the second
of three parts on speaker design by Larry Mundy.]
Efficiency
Ratings
Amplifier output
power and driver impedance are NOT the only things that determine
volume or sound-pressure levels coming out of your speakers.
There is another variable you need to understand, and that's
driver efficiency. A driver is like a pair of magnets. Two magnets
in proximity will either attract or repel depending on their
polarity. Attach a big paper cone to one of those magnets and
it will push air around. Reverse the polarity of the magnets
a certain number of times per second, and the pushed air will
reproduce a musical note at that frequency. That's basically
how a driver works; the alternating current from the amp, traveling
through the voice coil, sets up an electromagnetic field of constantly-changing
"polarity." The voice coil is suspended in a fixed
magnet which repels and attracts the voice coil according to
the input current, and the voice coil is attached to the cone
and pushes air around.
Not all of the
current passing through the voice coil is actually turned into
that useful electromagnetic field. Much of it, traveling through
the resistance of the voice coil wire, is turned into useless
heat. This is how your toaster (or indeed any electric heating
device) works; current passes through a bunch of thin wires and
heats them up. This is great for English muffins, but as we have
seen, it is not so good for drivers.
Driver efficiency,
in it simplest form, is a rating of how much of the current entering
the driver is turned into sound (and by extension, how much is
lost as heat), as well as how effective the rest of the driver
is at projecting that sound into the air in front of it. Because
so many driver-design variables determine efficiency, the accepted
rating system just focuses on sound output levels for a given
electrical input level, a real rubber-meets-the-road test. There
are three variables: the input power in watts, the sound output
measured in decibels ("db", think of this as "loudness")
and the distance from the driver at which this output is measured
(as you know from standing by the side of the highway, the resistance
of air greatly decreases the intensity of sound over distance).
The standard measurement is derived by feeding a driver one watt
of electrical power, and measuring the output in decibels one
meter from the front of the driver. This is sometimes called
"sensitivity," and sometimes "SPL" (for "sound
pressure level), and will be expressed like so: "SPL: 89db/1w/1m,"
or since the one-watt, one-meter measurement is fairly standard,
sometimes just "89db." Driver efficiency doesn't tend
to vary much with power input, so this one-watt test is just
as good as a hundred- or thousand-watt test for measuring efficiency.
You will see
such ratings both for raw drivers, and for finished, assembled
speaker cabinets. Ignore the latter for now. Cabinet designs
can improve (horn cabinets and ported boxes, at some frequencies
anyhow) or degrade (sealed boxes) the efficiency of a "raw"
or unboxed driver. No matter what you're going to put your driver
in, it just makes sense to look for a more efficient driver if
you can find one with the other attributes (power handling, frequency
response, physical size) that you want.
Most decent drivers
have efficiency ratings from the low 80's to maybe 110, although
it's very rare to see a driver designed for bass use approach
the magic 100 mark, and ratings over that are usually only achieved
by very efficient tweeters. So, how much difference is there
between a driver rated at 90 db, and one rated at 93 db? Well,
duh, 3 db, but how much is that really? Well, one electrical
way of looking at it is that it takes roughly twice as much amp
power output to drive the 90 db driver to the same "loudness"
as the 93 db driver. In terms of sound output, a 1dB difference
is the smallest change in sound level that is noticeable, and
a 10dB increase is perceived as "double" the volume,
so at a constant amp-current level, a 3 db increase is at least
"noticeably" louder.
The relationship
is "logarithmic" rather than "linear." A
doubling, or 10-db increase in volume, with a given driver, requires
10 times the amp power. The price difference between a 100-watt
amp and a 1,000-watt amp is generally hundreds of dollars. The
price difference between an 88db SPL driver and a 98 db SPL driver
can be literally zero. So you start to see the advantages of
shopping for drivers by efficiency ratings. Professional-audio
drivers (and cabinets) are designed with an eye to power-handling
and efficiency, to wring the highest SPL's out of a given amp/speaker
combination, even at the expense of some anomalies in frequency
response over their range. A good pro woofer should have an SPL
rating in the 90's. But within that general range, and all other
things being equal, you will like the 98 db driver a lot more
than the 91 db driver in terms of output. So when you're shopping
for drivers for live DJ use, selecting a more efficient driver
is almost like getting a free amp upgrade. Your amp doesn't have
to work as hard, but if it does, the result is louder sound.
What determines
driver efficiency? Well, for one thing, a light coil/cone structure
that's easy to move around has less physical resistance and can
be more efficient but if it's too flexible, it can flop
around and distort at high power levels, so few pro-audio drivers
seek efficiency that way. The same is true of very flexible surrounds
and spiders. That floppy little speaker in your TV can play pretty
loud with the 12 watt or 1 watt the TV amp feeds it, but fed
10 or 20 real watts it would sound fuzzy and eventually self-destruct.
So the search for efficiency in pro audio drivers generally focuses
on the voice coil magnet relationship.
Remember, the
real work is done by the interaction between the fixed magnetic
field of the driver's magnet, and the constantly-changing electromagnetic
field generated by the current through the voice coil. Since
that electromagnetic field is not really all that strong even
at high power levels, and decreases with distance as all magnetic
fields do, one way to increase efficiency is to lessen the distance
between the voice coil and the fixed magnet, sometimes called
the "voice coil gap." This is a matter of exacting
manufacturing tolerances, and is one reason for example that
JBL or Electro-Voice pro drivers are more efficient, and cost
far more, than equivalent Asian imports with the same magnet
weight, voice coil size and so on.
The most cost-effective
way to increase efficiency, with all other parameters optimized,
is to strengthen the fixed magnetic field the voice coil moves
in. Within certain limits imposed by the voice coil dimensions
and other physical parameters, a giant magnet can improve efficiency,
and even help recover some of the efficiency lost by a wider
voice coil gap. So more popularly-priced drivers sometimes attain
reasonable efficiency while retaining good power-handling capability,
by the simple expedient of adding magnet mass.
Usually this
is something touted in the specs: "Giant 30-ounce magnet!"
All other things being equal, I'd rather have a larger magnet
than a smaller one, because a larger magnet is usually a hallmark
of a better driver. A heavier-magnet driver usually costs more,
not so much because of the magnet (that stuff is fairly cheap)
but because the larger-magnet driver is usually a step up in
the manufacturer's food chain and also has a larger voice coil
or coil wires, or some other combination of other features that
allows the maker to charge more and still sleep at night. A heavier
magnet can allow the manufacturer to upgrade other important
components and still maintain a reasonable level of efficiency.
Other than cost, the only real downside is that you get to lift
those heavier magnets in and out of trucks and vans for the rest
of the life of the speaker. Watch out for a spec that touts a
heavy "magnet assembly," however. I don't know of a
standard that keeps a manufacturer from calling a non-magnetic
plate behind the actual magnet, the screws that hold it on, or
even some decorative doodad glued on the back, part of the magnet
"assembly." If the weight of the "magnet assembly"
is almost as much as the cited shipping weight of the entire
driver, there is likely some fudging going on.
Frequency
Response
Most driver specs
cite a frequency-response range, like "30 6,000 hz."
As a spec, this is sort of meaningless, because frequency response
can have dips and peaks, and if those dips in particular are
big enough, the driver might as well not have any response at
all at that point. It's better to look at a frequency-response
curve plotted under controlled and standardized conditions. If
the graph shows that the response at 44 hz (the low "E"
on a bass guitar) is 10 db less than the response at 440 hz (the
"A" an octave and a half above that), then that driver
is truly useless for bass, since the one note will be half as
loud as the other. Any fall-off of 3 db or more is significant,
and all drivers fall off at their frequency extremes. If a spec
cites "30-6,000 hz +/- 3 db," that tells you a little
more than the simple range, although still not as much as a graph.
To achieve high
efficiencies and power-handling, pro drivers tend to have rather
jaggedy frequency-response curves. What their overall usable
response rating is good for, however, is to judge whether the
driver will pretty much reproduce the range you want (does it
go low enough for bass guitar or synth? High enough for miked
cymbals?). Any decent DJ rig will have two-way or three-way cabinets
with woofers and tweeters and crossovers. The driver most likely
to be "blown" is the bass driver, because it has the
toughest job and DJ's tend to strive for bass "thump"
you can feel as well as hear. But if it will have to coexist
with an existing high-frequency driver, you need to make sure
that it has response that extends to the "crossover point"
between the two (the reverse is true if you are replacing a high-frequency
driver). Your cabinet manufacturer should be able to give you
this information. As a last resort, you can familiarize yourself
with crossover-circuit design (way beyond the scope of this article)
and work backwards to figure out the preset crossover point in
your speaker system.
Xmax/Xmech
One other spec
to glance at is "Xmax," which stands for "maximum
excursion" or more simply, how far the cone can move in
and out before reaching its limits while still receiving a signal
through the voice coil. "Xmech" refers to the mechanical
limits of cone excursion without regard to the voice coil signal,
and is a less useful (and usually larger) measurement. For deep
bass these can be useful things to know, because low frequencies
require moving large amounts of air, and a driver cone of a given
diameter can only move more air by increasing its cone excursion.
Bass drivers with lots of Xmax tend to have specially-designed
surrounds and voice coils that allow extreme movement, usually
at the cost of upper-end frequency response, efficiency, or both.
Again, it all depends on what sort of sound you're trying to
produce, or reproduce, but for DJ use, a bass driver with a lot
of cone excursion is usually the best choice.
More Specs
There are so
many other driver-design parameters and specifications that without
significant schooling and experience in electroacoustics, and
several pots of coffee to forestall the boredom, you might never
be able to choose between the gazillions of driver designs out
there. Many of the specs you will see (vas, fs, f3, "Q"
and so on) have more to do with how the driver will interact
with its surrounding enclosure, than how it will handle a high-power
input or how loud it will play with your amp. If you are replacing
a driver in an existing enclosure you can't change the enclosure
parameters, but you can check the driver manufacturer's suggestions
for enclosure size to make sure you're in the ballpark. If you're
really ambitious, there are a variety of computer-modeling programs
out there that will forecast very accurately the interaction
of a given driver in an enclosure of a certain volume, and with
ports of a certain diameter and depth.
Installation
and Testing
Let's say you've
bought a new driver and it's sitting in front of you. Find that
screwdriver, remove any grille over the blown or weak driver,
and clean off any "gunk" on the cabinet so your new
driver will have a flat surface for installation. The driver
terminals will be marked "+" and "-", or
sometimes the "+" side will just be marked with a red
dot. Hook it up the same way your old driver was hooked up. If
you don't remember, look for one of the wires to have a "hot"
color (in a red-and-black pair, for example, red is almost always
"+"), an identifying stripe in the "+" wire,
or even a raised ridge on one of the wires indicating "+".
Except in arrays of multiple speakers, you won't hurt anything
electrically if you hook the driver up backwards; it will just
operate "out of phase." Phasing problems will ruin
bass response in multiple-driver setups, because one cone is
moving in while the other is moving out, and vice-versa, so double-check
these connections. Make sure the wires can't wiggle loose inside
the cabinet, because you won't be able to fix that during a show.
Screw the new driver into the cabinet, making sure there are
no stripped threads or holes. The driver needs to be mounted
very securely, because it will spend its life trying to vibrate
itself loose. It's not a bad idea to check the driver-mount screws
or bolts in all your cabinets from time to time.
Test the cabinet
with your amp before you reassemble everything, to make sure
it works, then reassemble the cabinet. Now you should test the
amp-speaker combination more critically, because a different
driver may have a slightly different sound and require some small
tweaks in electronic equalization to achieve the sound you want.
Do this again after a show or two, because drivers tend to "break
in" during their initial use and become a bit more bass-heavy
as their cones become less resistant to movement.
Your sound has
reappeared, or perhaps even vastly improved. Even better, you
are now vastly more familiar with the "guts" of your
speaker cabinets. On to the next show!
Contact Larry
Mundy at larry.mundy@comcast.net
Next Month: Part
III
Related Links
Shavano Music Online
Speakerbuilding.com Interview with Joachim
Gerhard
AudioVideo 101
World Studio Group (WSG) Directory of
Related Sites
NAMM
Syn-Aud-Con
Audio
Engineering Society
Mackie
JBL
Pro
US
Speaker
Parts
Express
Industry
News
* New York's
DJ
Spooky
(aka Paul Miller) recently earned widespread acclaim for his
performance with the Oakland East Bay Symphony Orchestra on March
19, 2004. Composed by Boston's Anthony Paul Ritis, the concerto
for turntable entitled Devolution is based on the similarities
between the allegretto movement from Beethoven's Seventh Symphony
and loops employed by modern DJ performance artists. See San
Francisco Chronicle.
* Scottsdale
Community College and the Maricopa County Community College District
(the largest college district in the U.S.) formally approved
Club DJ as a two-credit Music elective on April 8, 2004. Disc
Jockey 101.com's Rob
Wegner
authored the official course competencies. The class will now
be called "Live Performance Disc Jockey Techniques"
(MUC 135) and is scheduled to be offered in the Fall 2004. This
marks the first time that a performance DJ course (not related
to Broadcasting) has been adopted for college/university credit
by an entire college district.
Tip of the Month