Loudspeaker Basics --
Part 3
While designers of loudspeakers can argue forever (and
will) over the perfect crossover design, those of us who are simply listening to their
products have a much simpler job of it -- all we have to do is choose a loudspeaker that
sounds good. If it sounds good, the why behind the design choice just doesn't seem
all that important. In this installment we'll glance at some differing design approaches,
but we'll primarily concentrate on how crossovers work. We'll let the designers argue over
the whys and wherefores.
In parts 1 and 2, we examined why crossovers are necessary:
a single driver cannot reproduce the entire frequency spectrum of music. Yet we
want a speaker to reproduce the entire sonic spectrum from deepest bass to the highest
notes produced in an orchestra, or as close to that range as possible. (A speaker that
actually reproduces that full range is called, not surprisingly, a full-range
loudspeaker.) Since no one driver is capable of re-producing that frequency range,
the musical signal must be divided into at least two parts. Then, two dedicated drivers
(tweeter and mid/woofer) can work together to cover a broader frequency spectrum
than each is capable of on its own.
A basic two-way monitor has a single crossover point.
It separates the tweeter’s frequency range from that of the mid/woofer. The exact
frequency of this point varies from maker to maker and model to model but tends to be
around 2kHz. To call this frequency the crossover point is a bit of a misnomer. To
illustrate why, let’s envision a piano. Each of its black and white keys represents a
tone. Let's divide the entire frequency range of the piano into three ranges: the bass,
the midrange and the treble registers. The bass register is to the left and the treble
register to the right. The middle register faces your (the player’s) chair. The very
lowest tone that the piano is capable of is obtained by depressing the key farthest to the
left. This also is the deepest dimension of the piano’s L-shaped body. It houses the longest
strings for the lowest notes. Just so, the right-most key on the piano’s
keyboard activates the hammer for the shortest string. It plays the very highest
note the piano can reproduce.
Let's look at an imaginary two-way speaker’s spec
sheet: It reveals that the speaker's frequency response is +/-3dB from 60Hz to 20,000Hz.
The crossover point is given as 2300Hz. Back at the piano, a key about halfway between
your seat at roughly the middle of the piano's keyboard and the piano’s right end
turns out to correspond to 2300Hz. Put a red tag on it that reads "crossover
point." Nobody could blame you for assuming that the tweeter reproduced all notes
from this particular red key on upwards. Using logic, you’d further believe that with
the red key, the mid/woofer took over. It’d cover all those keys left of it down
until its low-bass extension ran out. And why not? After all, the word crossover point
suggests exactly that: a very sharp separation, like a line drawn in sand -- the woofer on
the left side of the line, the tweeter on the right, and both drivers "butting
up" against it.
Suppose that were so, let’s think about the
requirements. What would be needed to accomplish such a sudden
"from-one-note-to-the-next" hand-off between two drivers?
In an earlier article, we already hinted at a
crossover’s operating principle. We called it a filter. A filter, in just about any
application, allows certain things to pass while others are blocked. In a speaker
crossover, an electronic component precedes the tweeter to pass the high notes and block
or filter out the lower ones. This part is called the high-pass filter. It usually
requires at least one capacitor. A low-pass filter (consisting of at least one inductor)
commonly precedes the mid/woofer. The low-pass filter passes the low notes to the
mid/woofer and blocks the high ones that the tweeter covers.
Let’s now play a series of chromatically ascending
scales on our piano. We’ll record this with a microphone and feed our signal via an
integrated amplifier directly into our speaker. We’ll start at the very bottom of the
piano’s keyboard, all the way over to the left. We’ll play each note with exactly
the same volume. We’ll be working our way up one white and black key at a time.
Actually, we'll press keys on the piano for some time
without hearing them over the loudspeaker. This is because the piano can reproduce tones
lower than our loudspeaker's 60Hz minus 3dB point. When we finally get up to about 40Hz,
however, or a bit higher (depending on background noise, room size, speaker positioning
and listening distance) we will begin to notice very faint sounds that slowly increase in
loudness as we continue to ascend the keyboard. These muted sounds are the significantly
attenuated (lowered) output of the woofer operating below its minus 3dB down point.
Even once we arrive at the key corresponding to 60Hz, the speaker’s output still
won’t be the same as the piano's. Though already significantly louder than
before, the speaker now plays at exactly half (50% or 3dB below) the piano’s volume.
This is just as the specification for bass extension predicted: - 3dB @ 60Hz.
Once our note-by-note ascent on the piano moves past
60Hz, we’ll begin to notice that the playback volume of the speaker now equals that
of the piano. Every note is reproduced at exactly the same level. In audio, the exact
point where this steady-state level begins is referred to as the point to which the
speaker’s bass extension is flat. Because our speaker is specified as already
being 3dB down at 60Hz, its point of flat bass extension of course must be somewhere above
60Hz.
Let’s continue our scale. Nothing surprising happens
for quite a while -- the mid/woofer continues to reproduce all of the piano’s notes
while the tweeter "just sits there" doing nothing. Then something unexpected
occurs. The more acute your hearing is, the sooner you’ll notice it: For a certain
span of notes on either side of the red key, both woofer and tweeter are active
together. In fact, you’d notice a mirror-imaged joint action. The woofer is
slowly fading out as you cross the red key. Simultaneously, the tweeter begins to become
louder. As you climb the scale past the red 2300Hz key, the woofer eventually becomes
inactive. The tweeter now takes over all the work. Its output level is once again
identical to your piano playing. If you happened to note where exactly the tweeter began
to work by itself, you’d know the lower frequency limit to which the tweeter is flat.
Now hold it. This cross-fading phenomenon wasn’t
exactly what we expected when thinking crossover point, was it? That’s
precisely why we called it a misnomer earlier. We actually need to think about it as a
crossover range. "Range" describes the area of overlap we
so clearly noted between both drivers. The reason for the overlap is simple. A filter
capable of fully passing one note and totally blocking the very next would
have to be exceedingly steep. That’s an audio term for a crossover network
filter that "kicks in" very sharply. It’s also called a high-order
filter.
The most drastic, steep or sudden high-order filter is
called a brick-wall filter. You may have heard this term in conjunction with the
upper frequency limit of the digital format. In the real world of engineering (rather than
theoretical ideals) a brick-wall filter is the closest thing to approaching the idea of
the crossover point (effective literally from one note to the next). A
low-order filter is more gradual or shallow. In loudspeakers with regular passive
crossovers, brick-wall filters are never used. Active digital crossovers are theoretically
capable of brick-wall severity but are still very rare.
First-order crossovers are the simplest, but require their
drivers to produce sound well beyond the crossover point, whereas third-order crossovers
do not place great demands upon their drivers, since each driver operates within a
specific frequency range and is abruptly crossed over.
|
