This week the team continued preparing the testing results for presentation. Writing of the first draft of the project report was also continued.
John
This week John continued drafting the report and organizing the testing results.
With three days of testing and over 100 data-points per day, proper organization of the data is vital to describing the testing results. Attenuation measurements were calculated by subtracting the dB SPL-C measurements without ANC from those with ANC.
Attenuation Measurements from Day 1 (Domestic Wall w/ Single Transducer)
Attenuation Measurements from Day 2 (Domestic Wall w/ 5 Transducers)
Attenuation Measurements from Day 3 (Window w/ Single Transducer)
Hunter
With the completion of the ANC test trials, calibration for the microcontroller’s (MC’s) SPL meter was attempted. To calibrate the MC’s SPL meter the MEM’s data sheet gives the microphones sensitivity (or analog output voltage to the input pressure ratio) as -45dBV/Pa under the given conditions: 1kHz, 94dB SPL, which gives 0.000511Vrms/Pa.
To calculate voltage using the dB equation we have:
Where the voltage being measured is the numerator, and Vref is the voltage reference level in the denominator. With the MEM’s sensitivity given as -45dBV, the SPL of for a dB measured at 40dB would be:
The (-2.8) is due to the pre-amp. For this project the pre-amp does not have unity gain and instead has a gain of 2.8, shown above, which is then subtracted from the SPL value giving the final result.
Unfortunately there were some major set backs that the ANC test trials required priority in order to complete on time. Although the testing preparations with the electret microphones applied the same method for the SPL calculations and received promising results, when attempting results from the MEMs microphone, making sure to recalibrate for the sensitivity of the MEMs microphone characteristics, the results were still far from expected in comparison to the electret.
This required immediate troubleshooting if there was going to be any attempt of completion. Hunter went to the west campus this past week to test any possible work around methods as this also meant the fundamental frequency was not achieveable either. This including making a full wave rectifier and precision rectifier, in attempt to change the AC signal to DC, since DC is usually easier for programing calculations.
The DC output seemed to have timing issues since a DC input was achieved but this only regulated the input voltage at whatever the initial DC voltage was.
Test Attempt for AC to DC Output by Applying Full-Wave Rectifier:
DC Input Displayed on Serial Monitor but only Regulating the Voltage:
Constrained with brute force programming as a final attempt, although this provided results that were even more accurate than the previous results, when testing with the electret microphone, yet still failed during the second attempt of testing with the ANC devices MEMs mic at last team meet up. If time wasn't a constrant, the teams feels confident that this would work with the ANC device.
Amplification using a LM4562 and LM386 was also attempted when the team meet up last, but the input was still low. This was very strange as the arduino mega's ADC needs a (5V/1024 = 4.8mV) change to increase the digital value where the MEMs does have an output higher than this and can only assume that the issue is either programming at this point or the possibility of it having something to do with its frequency ranged being focused on the lower frequencies. Overall, this is a major let down.
Since there seemed to be nothing more that could be done, with the amount of time remaining, the team decided to mount the touchscreen to the device to show that it can be turned from the device and that the buttons do interact, but does not display the expected values of the MEMs microphones due to limited testing. The ANC device's test trials, understandably, took the priority of time leaving the LCD's potential left only for the imagination.
