Sunteți pe pagina 1din 6

Xavier Bourret-Sicotte Physics

18/09/2007
Measuring the speed of sound

In this experiment, we will measure the speed of sound.

The apparatus consisted of a plastic tube filled with water linked to a water container. This
container could be displaced vertically in order to change the water level. We would then
make a tuning fork vibrate above the pipe and change the water level until the resonance was
at maximum intensity.
Hypothesis: The velocity of sound is 330 ms-1, the relationship between velocity, frequency

and wavelength is represented by the formula v=ƒ λ ,

1) Raw data presented

Table 1
Frequency /Hz Length /m
512.0 0.157
480.0 0.169
456.1 0.178
426.6 0.192
406.4 0.202
384.0 0.215
362.0 0.226
341.3 0.243
320.0 0.260
304.4 0.270
288.0 0.288
271.2 0.308
256.0 0.323
Error +/- 0.1 error +/- 0.001

Table 1 shows the frequency of the different tuning forks and the length of air needed for
maximum resonance. The error on the length was estimated to 0.1 cm as we repeated each
experiment several times. Note the error on the forks’ frequencies is a mere estimation as they
are accurately calibrated during manufacture.
2) Data analysis and interpretation
During the experiment, we have created a fundamental standing wave between the surface of
the water and the opening of the pipe. The distance measured represents ¼ of the complete
standing wave therefore λ = measured length * 4

-From the general formula v=ƒ λ , we find that velocity of sound in ms-1 for the first tuning
fork experimented is v = 512∗0.157∗4 = 321.536

-Calculating the error for velocity: ∂v = (∂ λ /λ + ∂ƒ/ƒ)∗v


For the first tuning fork: ∂v = (0.001 / 0.157+0.1/512) ∗321.536
∂v = 2.1108 ms-1
∂v≈ 2 ms-1
∴ v≈ 322 +/- 2 ms-1

-To find the mean velocity we add all values of v and divide the sum by the number of values.
A similar process was completed for the mean ∂v.

Results of all processed data are shown in table 2

Table 2
Frequency Hz-1 Length m-1 Velocity ms-1 ∂ velocity +/- ms-1
512.0 0.157 322 2
480.0 0.169 324 2
456.1 0.178 325 2
426.6 0.192 328 2
406.4 0.202 328 2
384.0 0.215 330 2
362.0 0.226 327 1
341.3 0.243 332 1
320.0 0.260 333 1
304.4 0.270 329 1
288.0 0.288 332 1
271.2 0.308 334 1
256.0 0.323 331 1
error +/- 0.1 error +/- 0.001

Mean velocity = (322+324+325+328…) / 13 ≈ 329


Mean ∂v = (2+2+2+2+2+2+1+1+1…)/13 ≈ +/- 2 ms-1
As v=ƒ λ , and ƒ=v/λ , let v be constant of 330 ms-1 then ƒ∝1/λ

By plotting the graph of Frequency over 1/λ we obtain a straight line. The gradient of this
line represents the constant v and the y-intercept a systematic error. The situation is illustrated
by graph 1

y = 313.24x + 16.736
Graph 1

550.0

500.0

450.0

400.0

350.0

Frequency Hz-1
300.0

250.0

200.0
0.55 0.75 0.95 1.15 1.35 1.55 1.75
1/wavelength m-1

Note that if we add the y-intercept to the gradient we obtain the value of 329.976 ms-1 for the
speed of sound.
Moreover, the second best fit line (in red) illustrates the most extreme interpretation of the
trend. Hence the difference between the gradients of the second best-fit and the best-fit lines
will represent the error estimation of the gradient of the graph.

Gradient best-fit line = 313.24


Gradient second best-fit line = 300.00
Error on gradient of the graph = 313.24 – 300.00 = 13.24
Gradient of the graph = 310 +/- 10 ms-1

3) Evaluation

By repeating the experiment using several forks of various frequencies we have


achieved an accurate measurement of the speed of sound. Indeed, although some estimations
of the velocity seemed far from the theoretical value of 330 ms-1, it appears that the average of
all measurements is 329 +/- 1 ms-1. The graph plotting frequency over 1/wavelength is a
straight line passing through all points within the error bars. This proves the inverse

proportionality “ƒ∝1/λ ”. Nevertheless, we notice that the graph’s gradient (313.24) and y-
intercept (16.736) do not equal the expected values of 330 and 0 respectively. Such distortions
are likely to be the consequence of an unknown systematic error yet because it has affected
the whole experiment in the same way, we can consider it as constant. If we add the constant
error expressed as the y-intercept to the estimated gradient we find the value of the speed of
sound to be 329.976 ms-1. Using the error estimation of the gradient found from the second
best fit line, our final result is v= 330 +/- 10ms-1
We may conclude that the experiment is a success as we have obtained the expected
values listed in our hypothesis within an acceptable error range.

4) Improving the investigation.

-The decision concerning whether or not the resonance was at its maximum was made
through human perception and is therefore subject to errors and misjudgements.
-In order to obtain a more accurate measurement of the resonance, we may have used an
electronic decibel-meter showing the “loudness” and therefore the amplitude of the standing
wave. Maximum amplitude means maximum resonance hence better readings and less
estimation.

-There has been a systematically occurring error not taken in account during the experiment.
Indeed, the measured length between the opening of the tube and the water level was
considered to be ¼ of the wavelength yet this is a mistake. The standing wave’s first antinode
actually takes place a small distance away from the tube’s end hence our readings were all
slightly biased.
- By taking this systematic error in account we can greatly reduce the difference between the
theoretical and experimental y-intercept of our graph.

-The size of the error on the calculated value of the speed of sound is inversely proportional to
the number of readings. By increasing the number of readings, we decrease the variation and
the influence of measurements exceeding the deviation.
- A larger data sample would allow a more confident discussion of the results and therefore
increase the accuracy of our conclusion.

- The plastic tube used to carry out the experiment was old and worn out. Moreover it was not
in a perfectly vertical position meaning the readings of the water level may have been altered.
Indeed, if the water makes an angle to the horizontal, the meniscus will be higher or lower
than it should producing a systematic reading error.
- By changing the tube and making sure it is vertically positioned during the experiment, we
will reduce such source of inaccuracy. This can be verified using a string and pendulum and
comparing its angle to the tube.

-Because the resonance was the consequence of the tuning fork’s vibration, we had limited
time to estimate the length of air needed for maximum amplitude of the standing wave. As the
tuning fork lost energy through dispersion of friction, it became difficult to take readings as
nothing could be heard anymore. We then had to hit the fork again and replace it at the top of
the tube. In order to take better measurements, we changed the water level within the tube
relatively fast. This will produce inaccuracies because our eyes will try to picture an instant
moment where the resonance is at maximum although the meniscus is in movement.
-By having several similar tuning forks ready to vibrate above the tube one after the other, we
can keep a constant source of resonance and have more time to take measurements. This will
decrease the error produced by the velocity of the meniscus.

Xavier Bourret Sicotte

S-ar putea să vă placă și