You are on page 1of 4

Physics 434 Autumn 2006

Module 3
Finding resonances in a Kundts Tube (lab due to and
thanks to Toby Burnett)
Introduction
This module explores the waveform capabilities of the NI DVI data acquisition board,
and the corresponding driver VIs. We will generate a sine wave of a given frequency,
use it to excite a physical system (the Kundts Tube), and measure the result.
We will also do some analysis of the results, demonstrating the capability of LabVIEW to
perform stand-alone data analysis.
The tubes that we use for this purpose have a loudspeaker in one end, and a small hole at
the other end for a microphone to measure the sound wave amplitude. There is also a
small box containing amplifiers for the speaker and microphone.
The point will be to find some of the resonances, and estimate the speed of sound, and
measure the damping.
We begin by considering the sound resonances of an ideal organ pipe, or thin tube, of
length L closed at both ends. Such a tube should show resonances under the following
conditions:
2L = n (1)
where L is the length of the tube, n is an integer, n = 1,2,3, ... and is the wavelength of
the sound. Standing waves are established such that at the ends of the tube, the pressure
amplitude is maximal. Such regions are known as (pressure) antinodes. As Figure 1
shows, there are also nodes (regions of the tube where the pressure amplitude is zero).
For n = 1, there is one pressure node at the center of the tube at L/2, while for n = 2, there
are two nodes at L/4 and 3/L4 and so on for higher values of n.
1
Physics 434 Autumn 2006
Figure 1. Standing Waves in a closed thin tube of Length L.
The wavelength of the sound wave, , is related to the speed of sound, c
s
and its
frequency of oscillation, f as follows:
c
s
= f. (2)
There are two methods for measuring the speed of sound in a resonant (Kundt's) tube. In
the first, the frequency of the sound wave is fixed and the length of the tube is varied
until a resonance condition is met. Then, knowing L and n, one can solve Equation 1 for
c
s
. In the second method, the length of the tube is fixed and the frequency of the sound
waves is varied systematically, producing a spectrum of transmitted sound pressure
amplitude as a function of frequency of the exciting sound wave. According to Equation
1, one will obtain a series of maxima in the transmitted sound spaced in frequency, , as
follows:
L
c
f
s
2

(3)
In practice, the spectrum does not consist of a series of infinitely sharp lines. Instead, due
to damping, an approximate model of the amplitude near line n is
( )
2 1
2
2 2
2
0 ,
1

'

,
_


+
f f
f f
Q A f A
n
n
n n n
,
(4)
where A
n
,
0
is the sound amplitude at the resonant peak and Q
n
is the damping quality
factor.
2
Physics 434 Autumn 2006
A block diagram of the swept excitation experiment is given in Figure 2. The excitation
signal is the digital to analog converter, which is used to produce a sinusoidal excitation
of varying frequency. The excitation signal is boosted by an audio amplifier which
converts the low voltage, low current signal to a high voltage high current signal suitable
for driving the speaker. The speaker transduces the electrical excitation signal to a
pressure wave. A voltage is applied to a voice coil (solenoid), which causes current to
flow, producing a magnetic field. This field produces motion of a rod inside the coil,
which in turn causes motion in the cone thus generating sound waves. The resonator is
the cylindrical tube in which the sound waves undergo constructive and destructive
interferences. These interferences manifest themselves as the distinctive pattern of peaks
(resonances) in the spectrum of sound transmitted by the resonator as a function of
frequency. The sound (pressure) waves are transduced to a voltage by a microphone. The
resulting voltage is converted to a digital value by the analog to digital converter.
Figure 2. Block Diagram of the Swept Excitation Experiment.
The raw signal from the sound experiment fluctuates at the frequency of the input sound
wave (carrier frequency). We are only interested in the part that varies as the frequency is
varied. The rate at which the frequency is varied in the experiment is much slower than
the rate at which the sound oscillates. We say that the amplitude variations are modulated
at the carrier frequency. We call this phenomenon amplitude modulation. Since we are
only interested in the amplitude of the output signal, the most straightforward approach is
to demodulate it electronically, producing a signal proportional to the amplitude.
3
Physics 434 Autumn 2006
However, to learn about signal acquisition, we will instead follow a more brute force
approach and analyze the signal digitally.
Procedure
Step 1: Analog exploration
This is an exploratory step that does not use the computer. Instead connect the signal
generator to the amplifier speaker input and the oscilloscope to the microphone output.
Set the waveform to be sinusoidal, and vary the frequency from 100 to 2000 Hz,
recording the approximate resonant frequencies, reading from the signal generator dial.
Make a quick plot of f
n
vs. n, and note that the spacing is linear only above ~ 800 Hz,
which differs from the nave expectation of Eqn (1). The rest of the lab will measure
these frequencies accurately. Leave the oscilloscope connected to the microphone output,
to monitor the following steps.
Step 2: Test the sound generator VI
Set up the sound generator VI to produce a fixed frequency, 1 kHz for example. You can
use a standard sweeping VI that is provided. Examine this on the oscilloscope, and verify
the frequency.
Step 3: Test the acquisition VI
You will use a capture VI to acquire the waveform. In this step, use the analog signal
generator to produce an input signal that can be acquired. For this, use the easy VI AI
Acquire Waveform. Then add a processing step to determine the root mean square (rms)
of the sine wave. (There is a VI to do this, of course!) This will be your measure of the
amplitude.
Step 4: Acquire the data
As you saw in Step 1, the linear part of f
n
vs. n is above 800 Hz. Set the sound generator
to sweep from 800 to 2000 Hz, and put your acquisition code or subVI into the for loop
sweeping over frequency. Acquire a table of rms amplitude vs. frequency and make a
graph and write it to a file.
Step 5: Fit the data
This is conveniently performed off line and will demonstrate LabVIEWs capability to
perform data analysis. You want to determine the best values of the parameters f
n
, and Q
n

for several of the resonances in the linear region. (You may have to edit the file to isolate
the individual resonances.) The library VI Nonlinear Lev-Mar Fit VI is designed for
this. Your output plot should show the data as open, unconnected circles and the model as
a solid line. (An example VI is provided as a starting point.)
Hand in your two VIs, from steps 4 and 5, fully commented and with documentation.
The VIs should be saved showing data and analysis results on the front panel.
4

You might also like