Last year, my physics class observed a demonstration that involved a Bunsen burner and a large sewer pipe, or a resonant chamber. When the resonant chamber was placed over the flame it made a loud sound. The demonstration involves a study of standing waves and beat frequencies, but it is unknown to the class as to why the resonant chamber makes the loud noise. Studying the nature of waves, I thought that the fundamental frequency might be a good place to start the investigation. Since the Tube had no holes in the side of it, the only series of tones that it can emit are multiples of its fundamental frequency (Giancoli 352). This would help to narrow down the possibilities from which the sound can come from.

However, the flame must be responsible for the sound to some extent, being that the tube does not emit a sound when just placing it into any space of air. Flames act in a erratic and strange manner. When ignited by a gas, such as natural gas, they stabilize by leaving the source and expanding as they rise, then they extinguish (See Diagram) (Flow and Combustion Simulation Laboratory 1). Upon further study I discovered that “a small gas flame introduced into a resonant chamber of air will emit a musical sound” and even is “practically the same pitch as the resonate note of the cavity” (Wood 212). Perhaps the very nature of the flame in its extinguishing and igniting rapidly causes the pitch.

What makes the flame cause the tube cavity to resonate however? Having observed fires in a wood burning stove at home, I have noticed that they often (almost exclusively) create a draft. Possibly this causes a change in the air pressure within the tube. Upon further research, I found that sound emits what is called sound pressure. This pressure is usually small and results from “alternating incremental changes in pressure that results from a sound” (Cornell Laboratory of Ornithology 2). In addition to this sound pressure, the air particles move together and apart as they reach the fundamental frequency, thus causing a change in pressure within a air chamber (Tyndall 61). This pressure could explain the noise that emits from the tube.

From observing the phenomenon and with a careful look at background information, I predict that the noise is emitted from the chamber when the frequency in which the flame ignites and extinguishes is equal to the fundamental frequency of the tube. Also, it appears that due to the aspects of sound and air pressure that the sound pressure within the tube changes and will be the greatest at the nodes (ends of the tube) and the least at the anti-node (in the exact center of the tube). If this proves to be true, than the graph of the pressure will form a sinusoidal wave as we take data at points moving down the tube.

Materials:
The materials that are needed to complete the experimental setup and the data gathering procedure are as follows: one 6 ft. long sewer pipe (preferably 8 in. diameter), one Bunsen burner, one drill with set of drill bits, one light meter, one microphone, one computer, an assistant to hold the sewer pipe above the Bunsen burner, electrical tape, masking tape, one tape measure, a natural gas source, matches or other means of lighting the natural gas, one piece of clear plastic, and one permanent marker.
Experimental Setup: (See Figure 1-1)

Before any data can be collected, a resonant chamber must be created and modified so that pressure readings and light readings can be made. First, drill 30 holes (1/2 in. or .0127 m in diameter) equidistantly spaced (using the tape measure) 2 in. (.0508 m) apart from the top of the tube. Second, cover all of the drilled holes with masking tape so they are airtight. Third, number the holes for future reference with the permanent marker. Fourth, drill out a 4 in (.1016 m) by 1 in. (.0254 m) hole where the fire will be in the tube and cover with a piece of clear plastic and tape it on with electrical tape to make a window. Make sure that the window is airtight. Then set up the bun son burner and put it onto the highest setting. Finally, plug in the microphone and light meter into the computer for data collection.

Data Gathering Procedure:
When collecting data, I took the light readings first as to avoid and blackening of the window, and thus a reduction in accuracy of the light meter data. To collect the data from the light meter, follow the following steps. First, plug the light meter into the computer. Second, turn the computer on and configure it for data collection with the light meter. Third, turn on the natural gas and then ignite the natural gas with the lighting device. Fourth, have your assistant place the resonant chamber (the sewer pipe) over the flame so that it starts to make the pitched sound. Fifth, look at the window to make sure that you can see the flame flickering in the window (Note: the flame should seem smaller when it is in the tube due to the decrease in the amount of oxygen that the flame is receiving). Once you have checked the flame, place the light meter against the window and then commence in collecting the data. Repeat the process as desired to obtain a set of data (Note: if the tube becomes too hot, stop data collection and let the tube cool off and then resume with data collection).

The pressure readings were subsequently taken using the following procedure. First, plug in the microphone. Second, turn on the computer and set up the program so that you are ready to collect the data (calibrate as necessary). Third, turn on the natural gas source and ignite the natural gas using the lighting device. Fourth, have the assistant place the sewer tube over the flame so that it starts to make the pitched sound. Now remove the tape from the top hole (hole number one) and place the microphone against the hole and take the data sample at that point. Once an accurate data point has been collected, remove the microphone from the hole and replace the tape over the hole. Repeat the process of placing the resonant chamber over the flame, removing the tape, taking data, and replacing the tape for all 30 of the holes, carefully recording data for each point (Note: if the tube becomes too hot, stop data collection and let the tube cool off and then resume with data collection).

Reasoning:
I chose this research setup because it seemed to be a reasonable way to test the phenomenon. The holes will allow simple data collection that should be fairly accurate. The window is placed on the tube since the flame should die down owing to the fact that there is less oxygen available in the tube than normally is offered for the flame. Overall, this setup seems to be a sound way to start an investigation.

Results

Sound Pressure Amplitude

Number (Hole From Top)	Amplitude (Low)	Amplitude (High)	Average Amplitude
1	3.656	2.83	3.243
2	2.684	2.772	2.728
3	3.538	3.099	3.3185
4	3.729	2.996	3.3625
5	3.812	3.548	3.68
6	3.846	3.534	3.69
7	3.822	3.597	3.7095
8	3.582	2.991	3.2865
9	3.763	3.065	3.414
10	3.817	3.661	3.739
11	3.871	3.573	3.722
12	3.758	3.734	3.746
13	3.792	3.28	3.536
14	3.792	3.148	3.47
15	3.778	3.719	3.7485
16	3.846	3.768	3.807
17	3.651	3.46	3.5555
18	3.861	3.846	3.8535
19	3.812	3.827	3.8195
20	3.778	3.773	3.7755
21	3.802	3.841	3.8215
22	3.895	3.846	3.8705
23	3.656	3.231	3.4435
24	3.9	3.646	3.773
25	3.846	3.661	3.7535
26	3.882	3.832	3.857
27	3.827	3.753	3.79
28	3.905	3.631	3.768
29	3.778	3.753	3.7655
30	3.817	3.363	3.59

Time (s)	Number of Periods	Amplitude of Potential (V)	Frequency (Hz)
0.002	1	0.471	65.79
0.0172	2	0.471	395.26
0.01973	3	0.471	104.06
0.02934	4	0.471	90.74
0.04036	5	0.471	105.04
0.04988	6	0.481	92.51
0.06069	7	0.476	207.90
0.0655	8	0.471	262.47
0.06931	9	0.476	116.14
0.07792	10	0.476	94.25
0.08853	11	0.476	97.85
0.09875	12	0.471	113.51
0.10756	13	0.476	n/a

Average Frequency	Frequency from Best Data	Fundamental Frequency of the Tube
(Hz)	(From 7th / 9th Periods) (Hz)	(Hz)

145.46	128.0409731	120.8442695

I found the uncertainty of my pressure data to be about ± .5 V from multiple trials on the same hole. The microphones uncertainty was negligible (.001 V), which proved to be consistent with what the manufacturer stated the uncertainty of the light meter and the microphone to be. Human error must have accounted for the greater majority of the uncertainty in the pressure data collection. For some reason (unknown to myself or my assistant) the microphone would read odd data if you help it against the hole at an angle opposed to square against the tube. The uncertainty of ± .5 V is rather large. It is not out of speculation that the pressure readings could support our hypotheses. However this seems unlikely, being that my assistant and I were consistent in our procedure, suggesting that our source for error should be approximately the same for each data point (and off in the same direction as well).

The light data was far more accurate. I found that uncertainty to be approximately ± .005V of potential. The peaks of the data were often difficult to discern, and the time difference error from ± .001 seconds. This problem could have created a problem with the calculated frequencies, but given that the data seems to support our hypothesis, I doubt that the data or calculations are seriously flawed, at least in the case of the light data. This made our frequencies skewed at some points, yet that could be just a speed problem of the light meter that should be accounted for, even though the manufacturer does not claim to have such a problem with it’s device. The cable that connected the light meter into the computer could also have contributed to a slow or decrease in performance of the light meter, thus effecting the data.

After observing the data, it is now clear that the pressure within the tube changes, but it does not appear to do so in a sinusoidal wave, which negates the final two parts of my hypothesis. However the pressure data that was collected was seriously flawed due to the large degree of human error with that aspect of the experiment. The graph of the amplitude of the pressure is clearly not in a sinusoidal wave, nor does it appear to have any correlation with any standing wave physics. The data is scattered and does not to seem to follow any set pattern whatsoever. Thus, the pressure within the tube does not appear to have any correlation with the sound that the tube makes. At the center of the tube there was no anti-node, nor was the pressure greatest at the ends of the tube. The air pressure that was being released from the tube during the investigation did not appear to change at all. The temperature of the air was the only part that appeared to change.

However, the frequency in which the flame flickers does appear to be closely related to the noise that the chamber makes. The fundamental frequency and the flickering frequency are approximately equal (128 Hz to 120 Hz). The small difference would fall within our uncertainty, so it is not unreasonable to think that the fire flickering causes the noise that the chamber gives off. On the other hand, I still cannot determine how the flame flickering at the same frequency causes the tube to resonate. Perhaps it might have something to do with the temperatures within the tube as well. The velocity of the air blowing out of the tube could be another area in which to investigate further. Whatever else that may be affecting the air chamber, it seems reasonable to conclude that the frequency in which the flame flickers is highly correlated with the noise that the tube emits. Another area that would be worth investigating would be the behavior of flames and whether or not they vibrate and/or cause the air around them to vibrate.

An interesting point to note is that the frequency of the pressure (113 Hz) was not far off from the fundamental frequency of the tube either. Perhaps this might have some correlation with why the chamber resonates in the manner that it does. Given the nature of the flame and how a flame flickers on and off. The pressure must increase and extinguish as the flame ignites and extinguishes. This seems to add further support to the hypothesis that the frequency of the fire is equivalent to the frequency the fire within the resonant chamber flickers on and off. This also suggests that, although the amplitude of the sound data that was measured with the microphone is suspect to much scrutiny, the frequency that the air was hitting the microphone was consistent with the frequency of the flame flicker. Thus the microphone data is not completely irrelevant, however it proved to support the hypothesis in an unforeseen manner.

Some places to improve the investigation are as follows: do not allow the chamber to become heated. That could have affected the results. By the end of my investigation, the intense temperatures had warped the test chamber, which could have contributed to error. Our equipment was also limited. The devices that were used could only take data at a certain rate, which made it difficult to calculate the frequency of the light flickering in the tube for example (and later the frequency of the pressure data). Case in point: use better equipment. Holding the chamber steady over the flame also became a problem at times, often extinguishing the flame all together, thus creating the need to run that test again and subduing the tube to more excessive amounts of heat. It would have also been nice to be able to gather both the pressure data and the light meter data at the same time. This was attempted, but it proved to be extremely difficult owing to the fact that the computer could not keep up with the massive amounts of data transfer. It was also difficult to even place the light meter and the microphone to their appropriate places at the same time while the assistant help the tube in place, all without being burned by the flame.

Fire is a curious phenomenon that often creates unusual outcomes. The problem with fire, as I found out with this experiment is that it is highly volatile and is difficult to control, even for a relatively brief period of time. The dangerous nature of fire is also a problem to be considered. Many safety precautions should be taken when working with fire, such as wearing gloves during the experiment (which was discovered by my assistant only after he burned his hands once holding the resonant chamber). This leads me to conclude that the fewer things that are required by a person to do, the better and the less error that can be involved.

Overall, the pressure data that was gathered was highly unreliable and did not seem to correlate with any known physics theory or evidence. Perhaps this could be because of the large amount of error in this portion of the investigation, but that seems unlikely, but possible. The light meter data seemed to work well. The setup worked well to serve its purpose and gather useful data. The frequency of the microphone data also seems to support a hypothesis that the fundamental frequency of the resonant chamber is equal to the rate in which the flame extinguishes and ignites within the resonant chamber. In future investigations, the holes used to measure the pressure could possibly be used for other means, such as temperature measures and velocity of the air leaving the chamber at known points. I am unsure as to what such data might imply, but it is worthy of further thought as well.

Given the analysis and the evidence that has presented itself, the following hypotheses seem reasonable to support and are worthy of further investigation from what has been observed. First, the frequency in which the flame ignites and extinguishes is high correlated with the fundamental frequency of the tube. Also, the velocity of the air inside of the tube might affect the chamber’s resonant nature. It appears as well that the flame inside of the chamber vibrates the air at the fundamental frequency of the chamber.

Works Cited

Flow and Combustion Simulation Library. Investigation of Flame Stabilization, Lift-off,

Table of Contents

Abstract Research Question Method Results Discussion Works Cited Back to research papers

Abstract

Research Question

Introduction

Results

Sound Pressure Amplitude

Works Cited