These are the first files in the file system to receive collected data. Each file contains one day's data. The file is created on by iMac when collection starts. The storage medium is Google Drive.

When midnight is detected, the file is closed and a new file is opened for the new day. The file is switched based on Linux local time, which means that files are switched based on Daylight Savings time in the summer.

The file name of these files contains the date when the file was opened. The format of the file name, with leading directories is:

FOMMX/data/ZinsByyyymmddAR.txt

where yyyy is year, mm is month from 01 to 12, and dd is day from 01. The “A” following dd is a sequence letter within day. If the collection is interrupted, this character is manually changed for a continuation file with the same date.

One day has 24*60*60 = 86,400 seconds. There are four or five samples per second. Therefore there are over 345,600 samples per file, or samples per day.

The data format in the files is that of Mathematica; the entire file is a list of lists. Each inner list is a sample except the beginning ones are comments. The beginning of the file is as follows

{
 {" program mainstore.py "},
{" File Name: /Users/stevenzins/Google Drive/gsync1/FOMMX/data/ZinsB20200126AR.txt "},
{" File Start Time: Sun Jan 26 00:00:00 2020 "},
{" Series Start Time: Sat Jan 25 15:56:15 2020 "},
{" type staq is Seconds, Turns, Azimuth, Quadrants "},
{" Seconds is since Linux Epoch"},
{" azimuth_of_house_axis = 333.2 "},
{"stdq", 1580022000.35, 484, 114.26, -1412},
{"stdq", 1580022000.58, 484, 115.58, -1411},
{"stdq", 1580022000.81, 484, 116.89, -1411},
 

These files are derived from the Original files. The data is resampled from about 4-5 samples per second in the Original files to exactly one sample per second in the Resampled file.

The data format in the files has also been changed to comma separated variables (CVS). While the Original files may have part(s) of a day missing, these files have only complete days.

The format of the file name, with leading directories is:

FOMMX/dataC/ZinsCyyyymmdd.csv

where yyyy is year, mm is month from 01 to 12, and dd is day from 01.

The file has CVS format and the first row is the field names. The first three records of file ZinsC20200126.csv are as follows:

timeMST,         turns,  azrotor, cumazrot,  quadrant
1580022000.81,   484.0,  117.0,   174357.0,  -1411.2
1580022001.8425, 484.0,  123.0,   174363.0,  -1410.0

• timeMST - This is the Linux time stamp. (Nothing to do with Mountain Standard Time)
• turns - Whole rotations of the rotor of the interferometer, derived from rotary encoder.
• azrotor - Azimuth of the rotor, in degrees, derived from rotary encoder.
  • Azimuth is zero when the interferometer arm with black tape is at true North.
  • Azimuth increases when the arm with black tape moves from North toward East.
• cumazrot - turns * 360 + azrotor
• quadrant - The interferometer determines the phase difference of the light from the two arms.
  • A phase difference of one quarter of a cycle, or of a wavelength, is one quadrant.
  • The quadrant value is the count of quadrants detected since the collection program on the interferometer rotor began execution.
  • A quadrant corresponds to one quarter of a full light-dark fringe shift cycle of mirror-style interferometers.

Each of the 151 Resampled files has been plotted as one page. This set of plots was produced to give a preliminary overall picture of the data, especially its major features. Centered above the plot is the file name (.csv) of the source Resampled data file. The upper left of the page has the file name of this page.

Each page has three frames. Each of the frames has the same time scale corresponding to hours 0 through 24 of that day.

The top frame has the Resampled data itself plotted twice but with two vertical scales. There are also two axes, on left and one right, The axis on the right has quadrant values from 0 to 7,000,000, which is the range of all plotted data in all Resampled files. This is the scale for absolute quadrant values on the chart; the line appears horizontal and straight because the variations within the day are very small compared to the full range of values.

The line that is not straight in this frame is the same data expanded vertically to fill the frame. The quadrant values of this line are those give by the scale at the left. The average of the values on the left scale should correspond to the position of the horizontal line against the scale on the right.

An example of one Plot Data File is for 2020-03-23.

The second and third frame each present the azimuth of the data found by a modified Fourier analysis. A normal Fourier analysis test for the phase and amplitude of a wide range of periods. Each of these two frames show a test for a single period. This approach is called the Goertzel algorithm.

The interferometer rotor rotates once a minute. The middle frame, labeled “Azimuth Full” uses the single period of one full turn. The bottom frame, labeled “Azimuth Half” uses the single period of one half turn.

Gallery of Plots of all Resampled Data

The collection period, described above, was January through June, 2020, about 180 calendar days. 151 Resampled Data files were generated with a full 24 hours of data. There are about 30 days where the Original Data files had less than a complete 24 hours of data and were not plotted.

Most of the gaps may have been due to utility power outages. Loss of power for a fraction of a second was enough to cause the iMac to stop requesting data or shut down. Loss of power for a very few seconds would also shut down the Raspberry Pi, with loss of the accumulated quadrant count. In both cases, manual intervention was required to resume data collection.

One large gap occurred from 20200417 through 20200428. This was due to the mechanical failure of the main thrust bearing that holds up the rotor and allows it to turn. The length of the down time was due mainly to the search for a better replacement part, the wait for delivery, and the installation of the replacement. While the collection was interrupted, I tried to examine the individual arm light output by connecting the fiber to an oscilloscope. In the process of doing this I inadvertently reversed the reconnections the two arms. The effect of this was to reverse the suspected Frame Drag component from -sin(24 hours) to +sin(24 hours). It probably also reversed the direction of the quadrant counter, but I am not certain of that.

The Original and Resampled data files are available upon request. The Original file set has has a little more data; the Resampled files have no gaps on any day. The Resampled files occupy about 4-4.6 MB each while the Original files occupy up to 17.9 MB.