DO'S AN D DON'TS OF G ROU N DI NG

J u n e 2 0 1 1
US $7.00/Canada $10.00

Tube, Solid State,

Loudspeaker Technology

Digital aX contains 45 pages, including: CES Report & Guide to Blu-Ray

SPEAKER TESTING
WITH DIY
ACCELEROMETER
IMPROVE
HOME
THEATER
SOUND
WITH LFE
MIXER

A QUICK
TURNTABLE
FIX

REVIEW OF OPPO BLU-RAY PLAYER

www.audioXpress.com
 speakers By George Danavaras

Accelerometer Testing of
Loudspeaker Drivers
Test your speakers' performance with this do-it-yourself measurement system.

I
n this article, I will describe the
design and operation of a system
consisting of the ACH-01 accel-
erometer and a preamplifier with
an integrated analog signal processor
(Photo 1). With this system, it will be
possible to measure at the low frequen-
cies the acceleration, the velocity, and
the excursion of a loudspeaker cone.
Also the panel vibrations of a loudspeak-
er cabinet will be very easily obtained.

The ACh-01-03
ACCeleromeTer
The ACH-01 (Photo 2) manufactured
by the Measurement Specialties (www.
meas-spec.com) is an inexpensive, gen-
eral-purpose, linear single axis acceler-
ometer with a very wideband response
(specified within 3dB from 2Hz to
20kHz) and internally buffered for low
output impedance. It can measure up to
±150g and has a high resonant frequency
at 35kHz. The output voltage of the ac-
celerometer is directly analogous to the
acceleration that it measures. The ACH-

PHOTO 1: The Accelerometer amplifier. PHOTO 2: The ACH-01 accelerometer.

6 audioXpress 6/11 www.audioXpress .com

Dana3278.indd 6 4/27/2011 2:39:35 PM

 FIGURE 1: Sine wave amplitude
definitions. FIGURE 2: The op amp integrator circuit.

01-03 version of the accelerometer is derivative of the displacement) and is and the RMS value of the displacement
housed in a small, rugged, flat package given by: is: Xrms = 3.53mm.
and is supplied with a shielded cable. The RMS value of the velocity is
a(t) = dv(t) = –ω2 × Xpeak × sin(ω × t) m and the RMS value of
Vrms = 0.444 sec
dt
Simple harmonic motion the acceleration is Arms = 55.83 m 2
sec
theory From the above relations you see that
Before I proceed to the description of the peak velocity of the cone is: Now what you need for your design is
the preamplifier, I think that it will be Vpeak = ω × Xpeak to apply the reverse procedure and com-
very helpful to refresh your knowledge pute the displacement and the velocity
about the theory of the simple harmonic And the peak acceleration is: of the simple harmonic motion when
motion. So here is a brief description. Apeak = –ω × Vpeak = –ω2 × Xpeak. the acceleration is known. Suppose that
Suppose that you have an ideal loud- the acceleration of a piston is a(t) as a
speaker cone that is driven with a si- The RMS values of the above are de- function of time t.
nusoidal signal as shown in Fig. 1. The fined as follows: (see also Fig. 1) Then by reversing the above, the ve-
instantaneous value of the displacement locity can be calculated by the following:
(=excursion) of the cone can be written For the displacement: v(t) = ∫a(t)dt
in the following form as a function of
time t: Xrms = Xpeak And similarly, the displacement can
x(t) = Xpeak × sin(ω × t) √2 be calculated by the following:
For the velocity: x(t) = ∫v(t)dt
Where Xpeak is the maximum value Vrms = ω × Xrms and
of the displacement and ω = 2 × p × ƒ is The above means that we can use an
the angular frequency. For the acceleration: integrator circuit to compute the veloc-
The velocity of the piston as a func- Arms = ω × Vrms = –ω2 × Xrms ity from the acceleration and then one
tion of time is the first derivative of the more integrator circuit to compute the
displacement and is given by: For example, assume that you have displacement from the velocity. This is
a loudspeaker cone, which is driven by exactly how the Analog processor of the
v(t) = dx(t) = ω × Xpeak × cos(ω × t) a sinusoidal signal with a frequency of accelerometer preamplifier operates.
dt
20Hz and the peak-to-peak displace-
Similarly, the acceleration is the first ment of the cone is 10mm. Then the the integrator circuit
derivative of the velocity (or the second peak value of the displacement is 5mm The integrator forms the basis of the ac-
celerometer preamplifier processor. Two
integrator circuits that are based on an
op amp are shown in Fig. 2. The left
is an ideal integrator and the right is a
more practical circuit. The main compo-
nent of these circuits is the operational
amplifier, configured in such a way that
its output voltage is proportional to the
integral of its input voltage.
The output of the circuit is given by
Vout = 1 ∫Vin × dt
RC
FIGURE 3: The block diagram of the Accelerometer preamplifier.
The resistor Rf reduces the gain of

8 audioXpress 6/11 www.audioXpress .com

Dana3278.indd 8 4/27/2011 2:39:48 PM

 5346 D

the circuit at the very low frequencies

(theoretically the gain of the ideal in-
tegrator at DC is infinite) because oth-
erwise the output of the op amp will
saturate after some time.
The integrator can also be seen as
an active low-pass filter with a very low
cutoff frequency.

The elecTronic Diagram

The preamplifier is based on the block
diagram of Fig. 3. The first stage buffers
the signal from the ACH accelerometer
and can provide a gain of 10 or 20dB
when the output of the accelerometer is
very low. The next stage is a high-pass
filter with a user selectable cutoff fre-
quency. An integrator follows and has
as output the velocity of the input while
the next integrator has as output the ex-
cursion (or displacement).
The complete electronic diagram of
the voltmeter is shown in Fig. 4. At the
input of the circuit, there is a three-
pin connector for the connection of the
ACH accelerometer to the amplifier.
One pin is used for the voltage supply
to the Accelerometer, the second is the
output of the accelerometer, and the
third is the Ground pin.
The voltage that supplies the acceler-
ometer should have a very low ripple and
for this reason a RC low pass filter consist-
ing of the resistor 24kΩ and the capacitor
470µF filters additionally the voltage.
Transistor Q1 with the components
R1, R2 and Led1 bias the internal FET
of the Accelerometer with a constant cur-
rent of 20µA. The jumper JP4 is used to
FIGURE 4: The electronic diagram of the Accelerometer preamplifier.
connect or to disconnect the bias circuit to
the input pins. This is helpful for the cali-
bration of the circuit, as you will see later.
The op amp IC1A and the compo-
nents around it buffer the output of
the accelerometer and provide with the
switch S1 a selectable gain of 0 or 10 or
20dB. It is very important that for the
measurement of the velocity and the
excursion the S1 switch should be set to
the 0dB gain (indicated as Normal posi-
tion on the front panel).
The components C13, C15, C16, and
R24 with the op amp IC1B form a first-
order high-pass filter and the switch S2
selects the cutoff frequency between 1,
10, or 50Hz. The op amp IC2A with
R3, R5, and C11 form the first integra-
FIGURE 5: The PCB layout of the Accelerometer preamplifier.

10 audioXpress 6/11 www.audioXpress .com

Dana3278.indd 10 4/27/2011 2:39:58 PM

 PHOTO 3: The PCB of the accelerometer amplifier.

FIGURE 6: The front panel of the Accel-

erometer preamplifier.

tor, which gives at its output the velocity.

The op amp IC2B with R6, R7, and
C12 form the second integrator that
gives as its output the excursion (or dis-
placement) of the input signal. Trimmer
R26 calibrates the output signal for the
acceleration, trimmer R22 calibrates the
output signal for the velocity and trim-
mer R23 calibrates the output signal for
the excursion. Switch S3 selects the type
of the signal that will be connected to
PHOTO 4: The placement of R11 and C17 on the bottom of the PCB.
the output of the preamplifier.
For the supply of the preamplifier I a limitation on the maximum dimen- shown. I printed this diagram to a
used an external power supply pack with sions of the PCB, which was not a prob- self-adhesive transparent paper, which
a nominal output voltage of 15V DC lem for this project. then was put on the front panel of the
and a current capability of 300mA. This The placement of the components on metallic box of the meter. This gave
gave an actual output voltage of about the PCB is shown in Fig. 5. a very attractive look for the meter as
16.4V DC but with a very low ripple All the components are placed on the shown in Photo 1.
because the total current consumption topside of the PCB except the resis- I also used the same diagram, print-
of the preamplifier is only 14mA. This is tor R11 (24kΩ) and the capacitor C17 ed on normal paper, as a guide for the
very important for the good operation of (470μF) that are mounted on the bot- opening of the holes for the rotary
the circuit. tom side of the PCB. Before the place- switches, the output connector, and the
ment of R11 the PCB track that con- power supply connector.
The PCB nects the power supply to the pin 3 of
The construction of the preampli- the Input connector should be cut. See CaliBraTion of The
fier is complicated so I designed a PCB Photo 4 for the details. amPlifier
using the Demo version of the Eagle The three rotary switches S1, S2, and The sensitivity S of the ACH-01 accel-
Layout editor. You can download this S3 are placed directly on the PCB to erometer is given in mV/g where g is the
m (the typical value
demo version of the program free from facilitate the construction of the meter. Earth’s Gravity in sec 2
m
Cadsoft (www.cadsoftusa.com). The In Fig. 6, the diagram of the front is g=9.81sec 2 ).
demo version is fully operational except plate that I used for the meter is Suppose that E (in mV RMS) is the

12 audioXpress 6/11 www.audioXpress .com

Dana3278.indd 12 4/27/2011 2:40:10 PM

 Madisound is pleased to offer the
Propeller Post speaker binding posts.
Propeller Posts are engineered by Seth
Krinsky of Virtue Audio. We have
been ogling these posts on the Virtue
Audio speakers and amplifiers for
years and we have finally convinced
Seth to let them out for distribution.
The Propeller
Posts are made
from Tellurium
Copper (99.5%
copper with
some Tellurium
making it easier
to machine).
The posts come
PHOTO 5: The setup for checking the accuracy of the system. in 16 and 35
mm lengths. If
you have space limitations, you can
order the posts with the Propeller Tool,
to allow for a
closer post
configuration.
Details and
pricing are
available on our
website.
madisound.com

Madisound is pleased to introduce

the Puresonic Spring Spade Terminal.
The fork in the spade is split to allow
them to be compressed when the
binding post is tightened. The spring
pushes back on the post and will not
loosen, assuring a very tight
connection.
PHOTO 6: Closeup of setup.

actual output voltage of the accelerom- by: xrms = Aωrms2 in meters. Now, for the
eter then the acceleration Arms that is calibration of the amplifier we will make
measured by the accelerometer is given some calculations based on the above
by the following: formulas.
The ACH-01 is provided calibrated
Arms = g × E m 2 from the factory and the one that I used The Puresonic spades are also made
S sec
had a sensitivity of S = 9.1mV/g. from Tellurium Copper and are a
If you know the acceleration and the If you suppose that the output of the perfect match to the Propeller Posts.
frequency (ω = 2 × p × ƒ), the velocity is accelerometer is E = 110mVrms at the The spades are available in either
computed by: frequency of 40Hz then you have: Rhodium or 24 carat gold plating.
The acceleration is:
Vrms = Aωrms sec
m
Arms = 9.81 × 110 m2= 118.58 m 2
9.1 sec sec
And the displacement is computed

audioXpress June 2011 13

Dana3278.indd 13 4/27/2011 2:40:25 PM

 I mounted the accelerometer on the
middle of the loudspeaker cone (Photo
6) using double-sided tape. The digital
caliper was firmly mounted perpendicu-
lar with the cone of the loudspeaker
using thick aluminum bars. The depth
probe of the caliper was moved down
until it touched the surface of the ac-
celerometer and the indication of the
digital caliper was set to zero (0.00mm).
Then the depth probe was moved
up some millimeters to allow the loud-
speaker cone to move and I increased
very slowly the voltage level to the loud-
speaker until I just heard the acceler-
ometer hit the edge of the depth probe.
This means that the positive peak excur-
sion of the loudspeaker cone was equal
FIGURE 7: Accelerations measurements.
to the indication of the digital caliper.
The measurements were performed
Table 1: Comparison between digital caliper indication for two different frequencies (15 and
and accelerometer amplifier measurements. 20Hz) and for three different peak cone
Digital Caliper Indication 5 mm Digital Caliper Indication 15 mm Digital Caliper Indication 10 mm excursions (5mm, 10mm, and 15mm).
Positive peak Negative peak Positive peak Negative peak Positive peak Negative peak The results are indicated in Table 1.
Frequency cone cone Frequency cone cone Frequency cone cone For the 5mm excursion at 15Hz
excursion excursion excursion excursion excursion excursion
(HZ) (mm) (mm) (HZ) (mm) (mm) (HZ) (mm) (mm) the indication of the preamplifier was
15 4.9 5.14 15 14.7 15.05 15 10.1 10.41 4.9mm for the positive excursion and
20 5.05 5.24 20 14.7 15.6 20 10.27 10.9 5.14 for the negative excursion.
At 20Hz the indication of the pream-
The velocity is: ity position and I adjusted the trimmer plifier was 5.05mm for the positive excur-
R22 so that the RMS Voltmeter indica- sion and 5.24 for the negative excursion.
Vrms = Arms m = 118.58 = 0.472 m tion was 47.2mV RMS (which means For the 10mm excursion at 15Hz
ω sec 251.32 sec m ).
that 1mV RMS = 0.01 sec the indication of the preamplifier was
And the excursion is: Finally I set the switch S3 to the Ex- 10.1mm for the positive excursion and
cursion position and adjusted the trim- 10.41 for the negative excursion.
Xrms = Arms = 118.58 = 0.1.877mm mer R23 so that the RMS Voltmeter At 20Hz the indication of the preampli-
ω2 63165.46 indication was 187.7mV RMS (which fier was 10.27mm for the positive excur-
Now you can proceed with the cali- means that 1mV RMS = 0.01mm). This sion and 10.9 for the negative excursion.
bration of the amplifier. For this, I used completes the calibration of the pream- For the 15mm excursion at 15Hz
a low distortion audio frequency gen- plifier. the indication of the preamplifier was
erator, a digital frequency meter, and 14.7mm for the positive excursion and
my true RMS digital Voltmeter (audio­ CheCking the SyStem 15.05 for the negative excursion.
Xpress, July 2009, p. 12). After the calibration of the preamplifier, At 20Hz the indication of the pream-
I set the frequency of the audio gen- I was thinking of a method to check the plifier was 14.7mm for the positive excur-
erator at exactly 40.00Hz and the level measurement accuracy of the acceler- sion and 15.6 for the negative excursion.
at 110.0mV RMS. I removed the jumper ometer and the calibrated preamplifier. The above results prove that the
JP4 (to disconnect the bias circuit) and Photo 5 shows the setup that I used: measurements of the accelerometer are
connected the generator to the input of A Peerless 830500, which is a 12″ very close to the indication of the digi-
the amplifier (pin 2 for the + and pin 3 loudspeaker. tal caliper.
for the – of the input connector). A power amplifier for driving the
I connected the RMS Voltmeter to loudspeaker. Some intereSting
the output of the preamplifier and I set A Fluke 289 True RMS multimeter meaSurementS
the switch S3 to the acceleration and I which can measure the peak (positive I used the accelerometer and the pream-
adjusted the trimmer R26 so that the and negative) values of a signal and plifier to perform some measurements on
Voltmeter indication was 118.6mV RMS A digital caliper. my current four-way loudspeaker system.
m ).
(which means that 1mV RMS = 1 sec 2 The demo version of the ARTA soft- Each channel consists of a separate closed
Then I set the switch S3 to the Veloc- ware (http://www.fesb.hr/~mateljan/arta/). box subwoofer with an Infinity 12″ speak-

14 audioXpress 6/11 www.audioXpress .com

 er which operates up to 62Hz, two Peer- supports the loudspeaker units to exam- to the loudspeaker constant at 20V RMS.
less 830452 woofers as open baffle dipoles ine their relative amplitudes. The drive I measured the distortion of the accelera-
that operate from 62 to 162Hz, a SEAS voltage to the loudspeaker was the same tion output signal at different frequencies
ER18 mid/woofer on an open baffle di- in all measurements. It seems from this and then the positive and negative peak
pole that operates from 162 to 2.2kHz measurement that from the 350Hz to excursion of the loudspeaker cone.
and a SEAS H1499 DXT tweeter. 500Hz the level of the front panel is For the measurements of Fig. 10, I
The measurements that I performed about –30dB down from the level of the increased the drive voltage to the loud-
are shown in Figs. 7-10. For the mea- loudspeaker cone. This indicates that speaker until the distortion of the cone
surements of Figs. 7 and 8, the loud- more damping of the front panel is nec- acceleration output voltage was about
speaker was driven with pink noise and essary for this range. 10%. Then I measured the drive voltage
the accelerometer was mounted in the In Fig. 8, I kept the same drive volt- to the loudspeaker and the peak-to-peak
center of the loudspeaker’s cones. I used age to the loudspeaker and measured the excursion of the loudspeaker cone.
the acceleration and the excursion out- excursion of the Peerless cone and the Unfortunately my test amplifier had
put of the preamplifier. excursion of the SEAS ER18 cone to a maximum output voltage of about
In Fig. 7, I measured the accelera- examine their relative amplitudes. For the 22V RMS, so it was not possible to
tion of the Peerless cone, the accelera- measurements of Fig. 9, I put the acceler- continue the measurements in higher
tion of the SEAS ER18 cone, and the ometer on the center of the cone of the frequencies that demand much more
acceleration of the front panel, which Peerless woofer and kept the drive voltage output voltage drive.

ConClusion
This is a very useful test and measurement
system. It measures the acceleration, the
velocity, and the excursion of a loudspeak-
er cone or a loudspeaker panel with good
accuracy. It is easy to use and the cost to
build is very reasonable. If you don’t have
such equipment in your lab, here is a good
opportunity to build one. aX

Parts list of the Accelerometer preamplifier

Part Value
C1, C11 1µ, MKT
C2, C3, C9, C17 470µ/25V, electrolytic
C4 22µ/25V bipolar
C5, C6 68n
C7, C14 100p
C8, C10 6.8µ/25V, electrolytic
FIGURE 8: Cone excursion measurements. C12 100n
C13 2.2µF/MKT
C15 33n
C16 150n
D1 1N4007
IC1, IC2 AD712
LED1 Green LED 5MM
Q1 BC548B
R1 4k7
R2 68k
R3 9.1k
R4 220k
R5 150k
FIGURE 9: Peerless 830452 measurements at 20V RMS. R6, R17 11k
R7 620k
R8, R9 27
R10 220
R11 24k
R12 1k5
R13 360
R14 3k3
R15, R16, R20 10k
R18 9k1
R19 5k1
R21, R24 100k
FIGURE 10: Peerless 830452 cone excursion and voltage drive for 10% cone accelera- R22, R23, R26 2k, Multi-turn trimmer
tion distortion. S1, S2, S3 Switch CK102X06

audioXpress June 2011 15

Dana3278.indd 15 4/27/2011 2:40:35 PM