|
CTS Piezoelectric Tweeters
Introduction
Piezo tweeters, in use for 30 years now, offer a quality
cost-effective, high-frequency sound source in a rugged,
high-efficiency package when used properly. Used improperly, however,
they fail to meet their potential. It is the purpose of this article to
help the user attain the maximum benefits from pie-zohs,
pizzas, pee-zoids, or whatever else they have
been called.
Background
Piezoelectricity was discovered by Jacque and Pierre Curie in the
late 1880s. They found that certain natural crystals generate an
electric field under the influence of a mechanical force. They named
the phenomenon piezoelectricity, from the Greek meaning
"pressure" electricity. The correct pronunciation is pi e' zo; however, Pe a' zo (the Latin pronunciation) has become as common.
Shortly thereafter, it was discovered that this phenomenon is a
reversible one. That is, when an electrical field is impressed across
the crystal, it undergoes a physical deformation. Since the actual
displacements are very small (measured in millionths of an inch), the
practical applications for piezoelectricity were slow in coming. The
various natural occurring materials were found to be piezoelectric,
among them quartz, tourmaline, Rochelle salt, and even wood. The advent
of radio resulted in the need for a frequency-stable circuit component.
quartz crystals, vibrating at resonance, were
found to operate consistently and are still the state of the art in
frequency-stable components. This was the first high-volume major
application of piezoelectricity.
Underwater warfare in W.W.II generated a need for detection
equipment analogous to radar used for planes. It was known that
acoustical signals travel extremely well in water, and the first
acoustical application for piezoelectricity emerged. A piezoelectric
crystal was acoustically coupled to the water through a metal
diaphragm. A short burst of energy (ping) caused the crystal to
vibrate, setting up and acoustical wave in the ocean. When the wave
encountered a hard object, a reflected signal was returned to the
sender. Since the piezoelectric device also worked as a receiver, after
the initial transmit "ping," it was switched to a receive
mode and listened for the returning signal. The time lapse between the transmit and receive was translated directly
into distance. Further, by adding multiple receivers aimed in different
directions, a direction (bearing) could be determined. Rochelle salt
was first used for this application because of its extremely high
sensitivity. Unfortunately, it exhibited several temperature and
moisture problems that made its use impractical.
A better material was needed. Independent research on both sides of
the ocean resulted in a family of synthetic materials that offer high
electro-mechanical conversion efficiency with greatly improved
temperature and humidity stability characteristics. This synthetic
material is actually a ceramic and is processed using methods similar
to conventional ceramic sintering techniques. The material is called
PZT because it is a polycrystalline lattice structure of the oxides of
Lead (P for Pb), Zirconium (Z), and Titanium
(T). Since it can be formed using conventional ceramic processes, it
offers more design latitude to the transducer engineer than do
crystals.
A major difference between PZT and piezo
materials found in nature (crystals) is that PZT must be processed
further to make it piezoelectric. The microscopic crystallites, known
as domains, are in random orientation in the PZT and must be aligned if
the material is to be useful. This is done in a process called
"poling." A high potential D.C. field is momentarily imposed
across the material causing the domains to align themselves with the
field. Upon removal of the field, the domains remain aligned (see
Figure 1). The poled PZT is now truly piezoelectric and will stay that
way unless an excessively high voltage is imposed upon it, or unless it
is heated to a very high temperature (Curie point). If either of these
conditions is reached, the energy input to the domains exceeds the
internal binding force holding the domains in alignment, and the
material once again becomes unpoled. This
entire process is very much like the magnetizing of a magnet except
that we deal with electric fields instead of magnetic fields. It should
be noted that the ability of the PZT to retain its polarity is a
function of the quality of the material. There are available low
quality materials which will de-pole under normal use causing the
speaker to gradually lose efficiency (sensitivity). CTS manufactures only the highest grades of PZT.
Theory of Operation
In operation, the domains within a poled PZT wafer (as shown in
Figure 2) alter their position slightly when an external field is
applied. This causes a slight deformation in the physical geometry of
the wafer. When the field is removed, the wafer returns to its original
size. These displacements are very small (measured in millionths of an
inch) but high in force, and when coupled directly to a liquid or solid
medium, are very useful for generating discrete motions. When coupled
to air, however, motions of these dimensions are useful only in the
ultrasonic region where the acoustic impedance of the air is higher,
and provides a better match to the PZT. To provide useful motion in the
audio region, a "mechanical lever", or transformer, is
required to convert the high-force, low-displacement motion to
low-force, high-displacement.

This is done by coupling two wafers face-to-face (Figure 3). The
wafers are connected such that as one expands, the other contracts.
When coupled at their faces with a metal member (centervane),
the resulting stress causes the sandwich to dish in and out depending
on the amplitude and polarity of the applied signal. This
"sandwich" is called a bimorph, as it consists of two active piezo elements.

|
By affixing a cone to the center of the bimorph and anchoring the
cone at its periphery, the bimorph vibrates in synchronism with an
applied audio signal and pumps the cone for and aft, while pushing
against its own mass (Figure 4). This concept, called the
"Momentum Drive Principle" was developed and patented by
Motorola in 1970. It is the fundamental principle behind a broad
family of speakers introduced in the ensuing 20 years through many
technical developments and dozens of patents.
|

|
Piezo Tweeter
Construction

Let's consider, in detail, the construction of the CTS Super Horn piezo tweeter. Although developed and patented in
the early 1970s, it is still a workhorse in commercial sound
installations. The circular PZT bimorph in this case consists of two
wafers, 0.89" in diameter and 0.0055" thick. The ultra-thin
wafer is required to achieve the desired acoustical performance. The
bimorph is coupled at it's center to the apex
of a specially impregnated diaphragm which then works into a
compression volume. Slots in the compression the compression space
direct the sound into the throat of the horn. The radial slots are
transformed into a 3" circular mouth through the unique shape in
the throat of the horn. The actual horn contour is a hybrid design
between a pure exponential contour and a hyperbolic one. Again, this
computer-generated geometry is optimized for the best acoustical
output.

The result is a frequency response (Figure 6) showing high
sensitivity and smooth characteristics. Again, it should be noted that
low quality products are available on the market using poorly tooled
parts and imprecise manufacturing methods. The results are inferior
performance and unpredictable results. CTS is
proud of its commitment to quality and the consistently high
performance of the full line at CTS piezoelectric speakers.
Piezo Tweeter
Performance
Because of the light dynamic mass of the piezo
tweeter (no voice coil, spider, etc.), the response is very fast. Tone
burst measurements show the excellent transient response at all
frequencies across the band.
A further advantage of the piezo tweeter
is its high-power efficiency. With no voice coil, there is no resistive
heating and little lost acoustical power. In fact, the actual impedance
of the tweeter (Figure 7) is very high, from about 50 ohms to 250 ohms
for the Super Horn in its operating range. At these values, the
amplifier sees the little additional load from the tweeter, allowing
use of arrays with little additional power load. Because it generates
little waste energy it is capable of being driven harder than the
dynamic tweeter.

Application
Hints
With all the aforementioned virtues of piezo
tweeters, there are still some issues in their use with which the
design engineer should be familiar.

The piezo tweeter appears like a lossy capacitor to the amplifier (Figure 8). As
shown in the impedance plot (Figure 7) the impedance decreases with
frequency. Many amplifiers today boast outputs that extend to 100 kHz.
At those frequencies, ultrasonic resonances may occur between the
amplifier and the tweeter, causing damage to one or the other or both.
If such an
amplifier is used, particularly with an array of tweeters, a small
series resistor is suggested (Figure 9). For CTS tweeters with a
low-end cutoff of 3 kHz to 6 kHz, a 50 ohm, 2 watt resistor wired in
series with each tweeter will prevent this resonance problem without
noticeably affecting the response. It should be noted that this problem
is uncommon in automotive applications since these amplifiers usually
roll off at 20 kHz. The 2 kHz horn products do not require an external
series resistor since one is built into each unit. The KSN1086
mid-range driver and KSN1090 and 1103 voice range products should be
protected with a 20 ohm, 10 watt series resistor.

Crossover Networks
|

|
The piezo tweeter does not require a
crossover network. Since the tweeter is capacitive in nature, it
rejects low-frequency power. However, if the mid-range is still
operating at the turn-on of the tweeter (4 kHz in the case of the
Super Horn), a harshness may be heard in the total system. This
disturbance in the crossover region can be minimized by the addition
of an R-C filter (Figure 10) tuned to attenuate the turn-on peak,
rolling off the mid-range a little earlier.
If a conventional crossover network is to be used, the tweeter
must be made to look "resistive" in order to work with the
crossover. This can be done by wiring an 8 ohm resistor /across/ the piezo tweeter. It should be noted, however, that
the power efficiency benefits are now lost since the piezo tweeter will look more like an 8 ohm
dynamic unit electrically. It will, however, allow the use of
conventional crossover technology. If a variable level attenuation is
desired, an L-Pad can be used.
|
|
If a straight level attenuation is desired, a simple (non-polar)
capacitor (Figure 11) can be series wired.
|

|
Multiple Tweeters
System sensitivity can be increased by adding piezo
tweeters in parallel (Figure 12). The high electrical impedance of CTS'
piezo tweeters allows several units to be
connected in parallel without overloading the amplifier.

Each time the number of tweeters connected in parallel is doubled,
the average sensitivity for the array increases by 3 to 6 dB. The
actual increase depends on factors such as off-axis angle, frequency,
tweeter model and the configuration of the array. For CTS Super Horns,
the on-axis response increases 6 dB for each doubling. Part of this
increase occurs because of the narrower beam produced by multiple
horns. As the beam becomes narrower, the off-axis response degrades as
a result of destructive interference between tweeters. The angle at
which the destructive interference is the greatest depends on the
frequency and on the spacing between the tweeters.
The destructive interference can be minimized in one plane by
orienting a single row or column perpendicular to that plane. For
example, if horizontal dispersion is more important than vertical, the
tweeters should be mounted in a single vertical column. This assures
that the horizontal dispersion of the array is identical to that of a
single tweeter. The vertical dispersion, however, can begin to degrade
significantly beyond 5 degrees. Mounting the tweeters at an angle
(off-axis) with respect to one another can also improve the off-axis
response.

Connecting piezoelectric tweeters in series doesn't increase system
sensitivity, but it does higher sound pressure levels at maximum rated
power (Figure 13). maximum power handling
capability of the array increases as tweeters are added in series. At
maximum rated drive level, doubling the number of tweeters in series
reduces the voltage across each tweeter by half with the resulting SPL
decrease of 6 dB for each tweeter. The additional tweeters, however,
create a 6 dB increase for a net on-axis sensitivity change of 0 dB. If
the voltage applied to the array is now doubled, so that each tweeter
sees it's maximum rated voltage, the array's
on-axis SPL increases 6 dB.


Power Handling
The power rating of CTS piezoelectric speakers is determined using
the EIA RS426 test method. This is a continuous 8 hour noise test with
peak voltage spikes twice (4 times higher in terms of power) the
average applied signal. Thus, for a speaker to be rated at 75 watts (25
volts), it must not degrade after 8 hours of continuous operation at 75
watts with 300 watt spikes. As a result of using the EIA test method,
CTS power ratings for its piezoelectric speakers tend to be
conservative compared to conventional industry claims for speaker
systems. In addition, the extremely dense, high-quality ceramic
manufactured by CTS withstands cracking and other high power failure
mechanisms much better than the piezoelectric ceramic used by many
other manufacturers.
Powerline Series
The Powerline series of 2 kHz horns use an
internal protection circuit which allows the horn to continuously
handle the full output of a 400 watt (8 ohm reference) amplifier.
The protector is a parallel combination of a miniature light bulb
and a positive temperature coefficient resistor (PTC)(Figure
16).

In a music system in which there is excessive clipping at high
power, or high-amplitude high-frequency signal content, the piezo drive element sees very large currents and
will heat up due to dissipation losses. When the PTC senses the high
temperature it increases its resistance dramatically. This has the
immediate effect of significantly lowering the power into the driver,
and the SPL produced. To avoid this sudden shift, and make the power
control practically imperceptible, the miniature lamp is wired in
parallel with the PTC. The lamp is essentially a very fact-acting PTC
and responds to music peaks rather than RMS heating as does the PTC.
The audible effect is similar to that produced by a level compressor.
In this way, the driver is held below damaging levels.
The resulting speaker performance then is as follows: under normal
operating conditions, the powerline speaker
performs in it's normal mode, faithfully
reproducing the signal applied in proportion to its volume. Under
temporary, extremely high power surges (even in excess of 400 watts), the speaker will still perform in its normal
expected mode. But now, if subjected to continuous high-frequency
power, above 100 watts or so, the PTC temporarily opens up, allowing
the speaker to continue to play, drawing its power through the light
bulb, at a somewhat compressed power level. The transition is smooth,
and at the power levels being played at the time, barely perceptible to
the human ear. When the speaker cools off, the PTC automatically
resets, and conditions return to normal.
Conclusion
The CTS product line of piezo tweeters has
grown dramatically since the Super Horn made its debut 18 years ago.
CTS' speaker portfolio now includes mid-range drivers, voice range products,
2 kHz horns, and the Power Line family.
CTS is committed to total customer
satisfaction. With the wide variety of models available, and the
technical tips provided herein, we are confident we can satisfy your
audio design needs.
|