Acoustics Lab
Acoustic lab: Does the data support the model?
Error analysis:
1.) How reproducible is your data?
2.) Please do a linear least squares fit on your data (resonant frequency vs n). To find the error in slope and intercept first calculate leastsqr = LinearModelFit[datalog, x, x]; then leastsqr["ParameterTable"] This yields the error in the slope and intercept. Plot the residuals to determine if the data support a linear model. Do the residuals “look” randomly placed about the origin (you don’t have to be quantitative here)?
3.) Is there a systematic error in the frequency calibration of your DFT phone app, which may lead to a offset in the fit for n=0?
4.) Did you determine the error in the model (working equation) due to errors in the pipe length and speed of sound? Does it account for any discrepancies in you data supporting the model?
Information on the model of acoustic resonators
https://en.wikipedia.org/wiki/Acoustic_resonance#Cylinders
Correction for the resonant frequency of a pipe due to end effects can be found here
https://en.wikipedia.org/wiki/End_correction
http://www.fonema.se/mouthcorr/mouthcorr.htm
Collecting and analyzing microphone data from the scope:
I want you to understand (1) the span of the DFT is half the sampling rate or samples per second and (2) the frequency resolution is determined by the duration over which you sample at this rate. For example, a sampling rate 2 Hz and duration 100 s, yields a frequency band from 0 to 1 Hz with 0.01 Hz resolution.
Please meet or email the TAs or me if you had trouble with the scope data.
I will open the lab if you want to collect more data. Just email to set up a time.
Lab report grading rubric, based on a 10 pt scale.
1. Writing (up to 5 pts off): Does the title make sense? Does the abstract indicate if the data did or did not support the model? Is there mention of how large the error is in relation to the model predictions in the abstract? Is the model described in sufficient detail to understand it? Does the body of the report make a logical argument that the data support (or do not support) the model? Is the conclusion appropriate? Are all the parameters used in the equations defined?
2. Citation (up to 2 pts off): Are equations and figures which are taken from other sources cited? Are some citations not from the web? Is the citation placed appropriately? Does the citation have a page number and date? Is the citation appropriate or just gratuitous?
3. Error analysis (up to 5 pts off): Is a working eqn given when appropriate? Does the error analysis follow from this working eqn.? Are there important systematic errors not discussed?
4. Data presentation (up to 3 pts off): Is the data presented in a easy to understand manner (no raw data)? Are the axes labeled on graphs. There should be no code in the report. Do the number of significant figures match the error or are there too many significant figures?
If you have questions, email or set up a time to talk with the TA’s or me. Our contact information is on the syllabus. We have answered LaTeX, Mathematica, and lab questions. We have also opened the lab outside of class time for students to make up labs.
Using Baudline
Click on the speaker icon in the upper tray Click sound settings On the output tab choose headphone On the input tab choose Analog input/microphone
right click on the display choose input and then scroll display to slow down the waterfall plot
right click on the display choose process and then transform size. The larger the number the smaller the span. 4096 works well. Notice the slider at the bottom to change the frequency span.
To generate noise go to playnoise.com and run the white noise generator out on the speakers or go to fourier sound 3 to run
