To ensure the signal sent to the wave maker was accurate, I ran over 40 tests to pinpoint errors. I looked to find reasons for the slight drift measured in the system and attempted to find the time step (and signal output resolution) best supported by the program and DAQ device. I found the drift to be a combination of signal noise and the small sampling output rate of the DAQ device. By taking a number of measurements, I was able to quantify fine travel adjustment factors for most usable modes of operation that counteract the drift effect. These adjustments are anywhere from .003v to .011v, but can be very important in long-term experiments.

In order to perfect the time step, I ran a large number of wave files which where created to send specific frequency waves at the probable time step. By reading free surface height with respect to time with a calibrated 1000Hz PXI card (running on a separate computer), I was able to use Fourier transport theorem to isolate the dominant frequency in the wave train and compare it to the intended frequency. With multiple iterations, I found the time step which accurately produced water waves to be .0094 (in the file writing program) and 9ms in the LabView real time iteration.

I also looked for motor power issues, such as variable results in different waver levels, but found that within the bounds of 0-25cm mean water height, the motor showed no signs of weakness.

Jordan Read (2006)