Experiment #1 Determining the Frequency Response Function of a Speaker by Your Name and John Judge1 Partner: Somebody Else ME 392 – Dynamics Laboratory Instructor: Dr. Joseph Vignola The Catholic University of America 02/??/2009 Include my name since you’re basing the report on the template I wrote. In the future, you will write your lab reports entirely yourself. 1 Abstract The Frequency Response Function (FRF) of a Somebrand Somemodel speaker was obtained experimentally. The speaker was driven through a Otherbrand Othermodel power amplifier with broadband white noise of amplitude ??? V, and the acoustic response ??? m directly in front of the speaker cone was recorded using a Differentbrand Another model microphone. Measured data was filtered using a 2-pole active lowpass filter with a ??? Hz cutoff frequency for anti-aliasing prior to analog to digital conversion. The FRF was calculated as the ratio of the frequency spectrum of the measured acoustic pressure over the input electrical signal. Results of ??? measurements lasting ??? seconds each were averaged to obtain a clean final result, which showed that the speaker has strong response between ?? and ??? Hz and had blah blah blah other distinct features. 2 Table of Contents2 Introduction ? Experimental Apparatus and Procedure ? Experimental Results and Discussions ? Conclusion and Recommendations ? References ? Appendix ? Note that I didn’t include either a “Nomenclature” or a “Theoretical Analysis” section in this lab report. The nomenclature section is necessary if there are a significant number of variables and symbols that must be defined for the reader – if there only a few, they can be described in the text. In general, theoretical analysis is needed for one of two reasons: (1) to generate theoretical predictions that you expect your experimental measurements to match, and/or (2) because the raw data is not meaningful by itself and further analysis must be done, based on some theory, in order to generate meaningful results. Neither of those is the case for this lab: the only processing applied to the raw data is a simple calculation of the FRF, which is a well-known procedure that does not need to be described in great detail. 2 3 1. Introduction Acoustic sources (speakers) have a variety of applications in research, industry, and commercial use. A typical speaker converts an electrical input signal into an acoustic signal in air or some other fluid, at frequencies within the audible range (20-20,000 Hz). More accurate transduction, especially for higher sound pressure levels, is usually accomplished by speakers optimized to perform over only a portion of that range, so that many audio systems are comprised of multiple speaker types (such as subwoofer, woofer, midrange, and tweeter) in order to accurately span the full audible range. A high quality acoustic source is one that responds with nearly constant amplitude and phase response to the electrical input single across a specific frequency range in which the speaker is designed to operate. This report describes an experiment to measure the frequency response function (FRF) of the Somebrand Somemodel speaker. The Somebrand Somemodel speaker is a ??-type speaker designed to produce sound in the frequency range ?? to ?? Hz. Add any other important description of the speaker here. The experiment entailed measuring the sound pressure level a fixed distance in front of the speaker cone, and processing the resulting data to obtain the amplitude ratio and relative phase between the acoustic signal and the input electrical signal used to drive the speaker, as functions of frequency. The experiment was controlled using a personal computer with a data-acquisition card, which generated the input signal used to drive the speaker, recorded the microphone output, and processed the resulting data. Section ? describes the experimental apparatus and procedure in detail, including the method for analysis of the measured data. The results are presented and discussed in Section ?, followed by conclusions about the effectiveness of the Somebrand Somemodel speaker in Section ?. Details of the custombuilt anti-aliasing filter used in the experiment are provided in an Appendix. 2. Experimental Apparatus and Procedure A schematic of the experimental setup is shown in Fig. ?3. The system is controlled by a PC running National Instruments LabVIEW software. A National Instruments PCI-6221 data acquisition card (DAQ), attached via shielded cable to a BNC-2120 connector block, allows the computer to pass signals to and from the experimental apparatus. The signal driving the speaker, which is generated by the PC, is first passed through a Otherbrand Othermodel power amplifier to provide sufficient current to the speaker. A Differentbrand Anothermodel microphone is mounted ?? m directly in front of the speaker cone. The microphone is powered by a PCB brand Somethingmodel ICP power supply (with unity gain), and the signal is passed through a custom built lowpass anti-aliasing filter prior to being acquired by the data acquisition card in the PC. 3 Make sure the figures are numbered in order and that any references in the text are to the correct number. Also, you can make the figures whatever size you like and integrate them into the text however you like – no need to use my empty boxes as exact size and placement rules. 4 Figure ?: Schematic of the experimental setup The anti-aliasing filter is a 2-pole Butterworth normalized active low-pass filter with a cutoff frequency of approximately ?? Hz [1], chosen to be significantly lower than the Nyquist frequency of the data acquisition system but higher than the expected response of the speaker. In addition to passive components, the filter includes a single JFET operational amplifier (LF351) powered by the built in power supply on the project board on which the filter was assembled. A circuit diagram for this filter, as well as details of its design are provided in the Appendix. Fig. ? shows the measured frequency spectrum of the filter, showing the cutoff frequency at ?? Hz and a ?? dB/decade rolloff at higher frequencies. Figure ?: Frequency response function of the custom-built antialiasing filter: (a) amplitude, and (b) phase. 5 For the results presented in this report, the drive signal delivered by the computer was uniform white noise with RMS amplitude 0.5 V, with a duration of ?? seconds. The signal was created digitally at a sample rate of ??? samples/second, and converted to an analog voltage signal by the output D/A converter of the DAQ. The power amplifier provides a fixed gain of ???, resulting in ??? VRMS delivered to the speaker (the speaker is rated for a maximum continuous voltage of ??? VRMS). The filtered signal from the microphone was digitized by the DAQ at a rate of ??? samples/second for ??? samples. Digitizer voltages were scaled by the microphone sensitivity of ??? Pa/V. The input and output were triggered from the same digital pulse so that excitation of the speaker and measurement of the response were synchronized. The experiment was run repeatedly a total of ?? times. For each iteration, the discrete Fourier transform (DFT) was taken of both the input and output signals, providing sets of discrete data describing the frequency spectra of each signal. The frequency response function (FRF), also known as the sinusoidal transfer function, is the ratio of the frequency spectrum of the system output to the frequency spectrum of system input, and is equivalent to the transfer function of the system assuming the response is entirely steady state and contains no transient behavior [2]. A discrete representation of the speaker’s FRF was obtained by dividing the DFT of the measured output by the DFT of the input . The ??? resulting functions were then averaged to obtain a signal result described the response of the speaker as a function of input frequency. 3. Experimental Results and Discussion Figure ?? shows the recorded time histories for a single instance of excitation and measurement. The same data are presented in frequency domain in Fig. ??, after applying LabView’s built-in FFT algorithm. The excitation amplitude is relatively uniform across all frequencies in the band of interest, while the measured sound pressure level varies as a function of frequency. Note that only the amplitude of the frequency domain data is shown in Fig. ?? (the complex numbers resulting from the discrete Fourier transform are output by LabView’s FFT algorithm in the form of amplitude and phase data). The Figure ?: Typical measured time histories of (a) excitation voltage send to speaker, and (b) sound pressure level recorded by microphone 6 corresponding calculated frequency response function is shown in Fig. ?? (both amplitude and phase of the FRF are shown). Note that the result contains some fluctuations due to noise. Figure ?? shows a less noisy representation of the frequency response function of the speaker, based on averaging the ?? functions for amplitude and phase obtained from each individual measurement. Due to the averaging, all portions of the measured FRF which are not coherent from one measurement to the next have been reduced by a factor of approximately ???. Describe the result in detail here, pointing out any interesting or important features. Comment on the extent to which you think the result is a pure representation of the dynamics of the speaker, or includes (unintentionally) dynamics due to the other components in the system. Justify any claims you make. Figure ?: Typical frequency spectrum of (a) excitation voltage send to speaker, and (b) sound pressure level recorded by microphone Figure ?: Typical frequency response function based on a single iteration of the experiment: (a) amplitude spectrum, and (b) phase spectrum. 7 Figure ?: Frequency response function determined by averaging results from ?? measurements, each of duration ??. 6. Conclusion The Frequency Response Function of the speaker over the tested frequency showed a strong response over the frequency range ??? Hz to ??? Hz, and was uniform in amplitude over that range to within a factor of ???. The response above and below that frequency range was (describe it). This indicates that the speaker should be used (describe any relevant conclusions). The experiment described here provides a measure of the speaker’s FRF using a minimal set of measurements using a simple laboratory setup. Describe any additional measurements or analyses that could be used to improve the results, or any recommendations to a future experimenter for improving the procedure. References/Bibliography [1] Scherz, P., Practical Electronics for inventors(1st edition), McGraw-Hill, New York, p.260, 2000. [2] Palm, System Dynamics, McGraw-Hill, New York, p. ???, 2005. Appendix Include a description of how the antialiasing filter was designed and constructed, with schematic diagram(s) and/or photograph(s) if necessary. 8