Experiments: Rotate Top or Floor Unit                          Latest change 2024-09-21
 
In brief:
To find out whether the alternating ellips is coupled to the structure of the pendulum the top unit and the floor unit have been rotated.

Conclusion: The phase of the periodic alternating is not coupled to construction parts of the pendulum.


Rotation of the TopUnit.
This experiemnt was suggested by prof. Schumacher. The aim was to confirm or deny a relation between the square structure of the TopUnit and the periodiic change in ellipse rotation, in sync with the N-S-E-W passages.


Fig 1. Result of 6 days data logging. (click for larger view)
The graph starts at 08-21_16:00 and runs to 08-27:10:00.
Yellow staircase: time of the day in hour steps,
Light blue: Foucault precession angle runs from +180 deg to -180 deg,
White: amplitude of the minor axis and the rotational direction of the ellipse. Positive is CCW, negative CW.
Data source: 2024-08-21.log .. 2024-08-27.log

Calculaton of FP
: We take only the zero crossing intervals where no rotation has been done:
from 23_04:47 to 23_18:41 gives 13:54.
from 23_18:41 to 24_09:27 gives 14:46, full period 28:40.
from 24_21:52 to 25_13:54 gives 16:02
from 25_13:54 to 26_06:21 gives 16:27, full period 32:29.

Actions:
Up to 08-21_15:00 the situation was as I call it "normal".
On 08-22_15:00 I rotated the top unit 15 deg CW. We see the blue line go up from 81 to 96 deg. (exact numbers read from the log files)
I could also see a substantial shift of the plot on the compass display, but that was corrected for in a few minutes by the autozero mechanism.
At 16:00 I did relaunches in NS and EW to see if the center timing error had not became to large. I decided the error to be acceptable. The last relaunch in these actions was done in near E-W direction.
On 08-24_10:00 I rotated the topunit 37 deg in CCW direction, to the other side of the "normal" position. We see the blue line go down from 158 to 121 deg.
On 08-26_17:30 I rotated the topunit back to its "normal" position. We see the blue line go up again from -117 to -93 deg.

Analysis:
When the topunit is rotated we see only an effect on the measured precession angle. That is to be expected, the angle measuring devices (Hall sensors) sit on the topunit.
We see no jumps in the measured length of the minor ellipse axis and its rotational direction which can be related to the rotation of the topunit. This means that the synchronisation to the N-S-E-W directions is stil present.
We do see some change in the linearity of the FP and in the duration of the positive and negative parts of the ellipse rotation, particularly from 08-24 to 08-26. So there is something else which affects the regularity of the FP. A suspect is the wire, which is known to be slightly bended. See TopMount, fig. 6.
We also see at certain places a small sudden jump from positive to negative (not the other way) in the ellipse amplitude. I regard this as an artifact caused by certain small timing errors in the measured and processed data. It does have my attention.
We can detrmine the FP period during the full FP rotations without intervention.
That is from 08-23_04:46 to 08-24_09:27 or 28:51. So to fast.
And from 08-24_21:51 to 08-26_06:21 or 30:30. So very accurate for NL.

Conclusions and discussion:
The periodic change of the ellipse's rotational direction is not related to the construction of the TopUnit.
Remaining candidates are: instability of the way the top unit is mounted to that beam of the roof, and that beam, Earth Magnetic field in combination with the magnet in the bob or magnet on the top, Earth Rotation.
Roof: that beam points some 15 deg west from the north. Not very likely that it will wobble exactly N-S or E-W.
Geo Magnetic Field (GMF): Earlier I have seen a serious effect from a magnet in the vicinity of the bob, having there a horizontal fieldstrength comparable with that of the earth. And in NL the horizontal component of the GMF coincides quite well with the geographical north, so it is nearly indistinguishable.
While writing this another test came to my mind: reverse the magnet in the bob and see what happens. Will the ellipse direction change? I did the test immediately and below is the result: negative, no change.
I plan to construct a large coil set, such that I can manipulate the magnetic field in the vicinity of the bob.
The posibility of the anisotropy in the wire will be investigated by rotating only the proxxon clamp with the wire. I then also will have a carefull look at the bob, It should rotate as much, but it's worth to verify that.


Floor unit rotated approximately 30 degrees CW:


Fig. 1 Floor unit rotated CW.
The floor tiling in my house points ca 15 deg west from the north, so the vertical in this picture is almost N-S.
and the floor unit is pointing ca. 15 deg east from the north.
In all previous experiments the floor unit was aligned with the tiling.


Fig 2.  Rotation of floor unit CW.  
Legend:
Yellow staircase: Time of the day in hour steps
Light blue: Angle of foucault plane +180 deg .. - 180 deg
White: Amplitude of minor axis. + = CCW, - = CW, 45 mm / div
Graph starts at 2024-09-08 at 01:00 and runs to 2024-09-09-12:10.
Zero crossing intervals of FP:
08_01:28 to 08_16;59 gives 15:31.
08_16:59 to 09_08:57 gives 15:58, full FP was 31;29.
Data source:  2024-09-08.log .. 2024-09-09.log

After this experiment it was discovered that the top unit had not been rotated back to the original position, aligned N-S.
This has been redone befoer the next test, floor unit rotated CCW


Floor unit rotated approximately 30 degrees CCW:


Fig 3  Floor unit rotated CCW
The unit is now pointing ca. 35 deg west from the north.


Fig. 4 Rotation of floor unit CCW.
Legend:
Yellow staircase: Time of the day in hour steps
Light blue: Angle of foucault plane +180 deg .. - 180 deg
White: Amplitude of minor axis. + = CCW, - = CW, 45 mm / div
Graph starts at 2024-09-11 at 11:00.
Zero crossing intervals of FP:
11_11:18 to 12_02:12 gives 14:54.
12_02:12 to 12_16:52 gives 14:41, full FP was 29:47.
Data source 2024-09-11.log .. 2024-09-13.log

Conclusion: The phase of the periodic alternating is not coupled to construction parts of the pendulum.

Note: in all pictures of the minor axis amplitude irregularities are visible. The origin is not known, however air flow is stil a suspect.
In near future Iĺl construct a windscreen.