wind_sensor:meeting_minutes_mar_3_2017

Attended: Mengyuan, Creighton, Scott

  • Updates:
    • Conducted additional tests to verify accuracy of moving average algorithm
      • Results are slightly inconsistent. For a 4 mph wind speed, the output of the code is on average 6 +/- 1 mph.
      • Adding foam padding around the microphone improved results
        • Detected wider range of frequencies
        • Results improved to 4.5 +/- 0.1 mph
    • Worked on Preliminary Design Review
  • To Do:
    • Recalculate coefficients with padded microphones to see if previously determined coefficients change
    • Conduct additional tests
  • Updates:
    • Completed the program that saves the clock cycles in an array and prints out the result at the end. Our results were okay half the time and garbage the other time.
      • In trying to debug the problem, we noticed that there was sometimes a problem with the array being lined up properly. If a microphone triggers twice in succession, it may create two entries instead of one, which will offset that array and will cause all the differences to be skewed. We noticed that this may cause mic2 to appear to have been triggered before mic1, which results in a negative difference. This may be due to a lack of hysteresis - which is the detection of a signal going from HIGH to LOW. Because we're just using voltage comparators to detect when it is triggered, a signal hovering around the threshold may cause it to trigger falsely. A Schmitt trigger is a device that can achieve hysteresis.
      • Along with that problem, we were a little worried with our results for when it does work. Recall that we had to get within 0.3us for this method to be effective. However, we were seeing results that vary on the order of microseconds, which is too large. One possibility is that the interrupts may actually be blocking (i.e. mic1's interrupt prevents mic2's interrupt from triggering on time), which may alter our results.
    • Received the Schmitt Trigger and transducers. Unfortunately do to an oversight on our part, we realized our Transducers require an AC signal to be powered, instead of a DC signal as we thought. However, we think we might be able to use the on-board DAC to create a 40kHz sine wave to drive it.
    • Worked on the presentation.
  • To Do:
    • To check if our interrupts are actually blocking, we can do a simple test where we look at the differences in activation times for the mics.
      • For example, if mic1 was triggered at times 0s, 1s, 2s, exactly, then that's a good sign since it's a constant 1s. However, if instead we see that it was triggered at times 0s, 1s, 8s, then we know that mic2 might have been blocking mic1 from activating.
    • Lots of tests:
      • Look into using high speed ADC found here: https://forum.pjrc.com/threads/25532-ADC-library-update-now-with-support-for-Teensy-3-1. Only use a limited bit precision to try and get the highest speed.
      • Measure the clock cycles involved with outputting a 40kHz sine wave from the DAC. If it's not too much, it may be feasible to drive it from the Teensy.
      • Run a test anyway using a clock cycle counter to measure the time from the emitted signal (start of DAC) and the received signal in the receiver.
      • One possible alternative is to use thresholding to detect a wave, and then run some ADCs to determine the phase shift afterwards.

Authors

Contributing authors:

mwu snishihara

Created by mwu on 2017/03/02 18:45.

  • wind_sensor/meeting_minutes_mar_3_2017.txt
  • Last modified: 2017/03/04 03:33
  • by snishihara