Assess your understanding of Signal Processing
  1. The peak signal to rms quantisation noise for a 16 bit ADC is:
    (a) 122dB      (b) 78dB      (c) 107dB      (d) 95dB

  2. The required filter to combat spectral aliasing is:
    (a) Low pass      (b) High pass      (c) Bandpass      (d) Notch

  3. The Fast Fourier Transform assumes that the signal is
    (a) Periodic within the time block of the measurement window.
    (b) Periodic within the measured number of averages.
    (c) Stationary within the resolution bandwidth.

  4. For n un-windowed ensemble averages the standard error of estimate for each spectra line is:
    (a) 1/n     (b) 1/√n     (c) 2/n     (d) √(2/n)

  5. The amplitude unit appropriate to the power spectral density function of an accelerometer signal is:
    (a) g     (b) g/Hz      (c) g2/Hz     (d) √(g2/Hz)     (e) g2/√Hz

  6. You are required to analyse a signal recorded from a random vibration test using a 2048 point FFT analyser (1024 lines). The analysis will require 50 non-overlapping ensemble averages and the spectra will have a Nyquist frequency of 20 Hertz. A test engineer consults you for advice on the analysis duration. Would you recommend:
    (a) 2.5 sec      (b) 25 sec      (c) 256 sec     (d) 2560 sec

  7. Dynamic engineering signals can be displayed as functions of time or of frequency, the two being related by the Fourier Transform. In each of the following cases select the figure on the right which most closely corresponds to the signal on the left.