|
Comparison
methods for testing frequency response of loudspeaker under anechoic
chamber free condition. In
principle, the frequency response of a loudspeaker should be measured
in an anechoic chamber. The anechoic chamber can (1) isolating noise
from outside; and (2) eliminating reflection waves. Because
to construct an anechoic chamber is very expensive, the following
three anechoic chamber-free methods have been developed : (1)
TDS method; (2) pulse FFT method; and (3) Special test box method
(developed by SUNLIGHT). Amount
them, the TDS method is obsolete. The principle will not be described
here any more because now this method is seldom used. Here the methods
(2) and (3), which are common used nowadays, are compared. All
other electro-acoustic measuring instruments use the pulse FFT method.
It claims that anechoic chamber may not be used. However, if
measurement is made under free anechoic chamber condition, a lot of
disadvantages comes : A.
The pure test space should be so large than not practical. B. Operator should have deeper technical knowledge. He is able to make accurate calculation on
the reflecting time. These are the basis for correct operation.
C. The frequency intervals at low frequencies are too large. For midrange and woofer loudspeakers, the intervals are larger than practical requirement. It is only acceptable for the measurement of
tweeter .
D.
Distortion can not be measured under free anechoic chamber
condition. E. Results are not stable because it is easy to be influenced by the environment noise. Results may be
different even under very low background noise.
However,
SUNLIGHT 1600F has not the above shortcomings. The special design of
1600F is originated from SUNLIGHT. This one of the inventions from
SUNLIGHT. The
following description explains why the pulse FFT method has above five
shortcomings. In
this method, a very short voltage pulse is fed to the loudspeaker by
the instrument. The propagation time of the direct wave of the pulsed
sound from Points P to Q is about 2.9 ms (0.0029 second). The
propagation time of the reflected wave of the pulsed sound from Points
P to M to Q is about 9.2 ms. In order to avoid the effect of the
reflected wave, the sampling time of the A/D converter must be chosen
as 6.3 ms (see lower figure). Then FFT calculations are made to derive
the frequency response curve of the loudspeaker.
According to the FFT theory, the frequency interval should be the reciprocal of the sampling time 6.3 ms, i.e. 160 Hz. It means that the measured points which form the curve have equal frequency interval of 160 Hz. What we know is only limited to the frequencies of 160 Hz, 320 Hz, 480 Hz, 640 Hz, 800 Hz, 960 Hz... Nobody knows the response below 160 Hz, and at frequencies between the neighboring points. Exactly to say, the given curve is a smoothed segment. To
reduce the frequency interval, the distance of the shortest reflection
path (P to M to Q) should be increased.
However, this distance is limited by the real size of the room,
so that it is not practical in most cases. In other words, the
interval is very difficult to be smaller than 160 Hz. We
can conclude that this method is suitable only for the measurement of
frequent response at higher frequencies (e.g. > 1 kHz). Easy
to be influenced by noise : When the
pulse FFT method is used, the pulse fed to the loudspeaker is very
narrow, that means the “Pulse Energy” is very small or the
“Pulse sound energy” generated by the loudspeaker is also very
low. This is why the measured frequency response curve is not accurate
under noisy environment. In
SUNLIGHT 1600F, entirely different principle is used to reach the free
anechoic chamber measurement.
There is no above shortcomings. A special designed test box is used in
1600F. The box isolates environment noise considerably. With the
addition of the “Tracking Filter”, the influence of environment
noise is further eliminated. These measures ensure that 1600F can be
used in noisy production line. The
Test Box 143B in 1600F has special construction. Together with the
software corresponding to the construction of the box, the free
anechoic chamber test
system is formed. Since the characteristics of individual Test Box
could not be exactly the same (e.g. slight difference in the density
of the sound absorbing material), each 143B must be calibrated and
corrected against a large anechoic chamber individually. The
calibrated result is stored in the software so that the disk of one
143B can not be used in other 143B. In other words, each 143B must use
its own attached software. It is true even for two 143Bs manufactured
in the same batch. The
frequency response (including sound pressure sensitivity) of a
loudspeaker measured by SUNLIGHT 1600F is the same (or very close) to
that measured in a large anechoic chamber according to CNS 4784 or JIS
5531 National Standard.
|
