Reproducing the Velador Experiment

Conclusions and Additional Speculations as of 6/19/07

I have adopted an assumption of dependence on sidereal time for these reasons:

1. I have mounting experimental evidence that the peak-to-trough amplitude of the measured effect is time variable with a period of roughly one day.

2. Dependence on anything other than sidereal time with a period of one day implies dependence on local phenomena.  (This does not exclude phenomena acting on the laser light, such as the Sagnac effect, but does rule out any absolute vector reference.)  The measured effect is large enough that a sidereal time dependence can be differentiated from a diurnal temperature dependence.

3. My current experimental plan provides no means of differentiating between thermal drift and any known effect not dependent on sidereal time.   Sidereal time dependence can be isolated because the locally measured effect will change predictably over the course of the year, while an effect dependent on the solar day will not.

Dependence on sidereal time is the only explanation remaining for some of my experimental results that does not preclude an absolute vector reference.  So, I have decided to make sidereal time independence a rejection criterion.  Because sidereal time is essentially a measure of astronomical position, this implies that results are dependent on orientation, not time of day. 

Other observations to date include the following:

1. Inherent image noise is slightly smaller than anticipated – approximately 1.5 pixels actual peak-to-trough noise vs. up to 3 pixels anticipated noise.  This makes the Velador slightly more sensitive than anticipated.

2. There is a “node” of minimum peak-to-trough amplitude observable in the early morning at this time of year.  This is coincident with 0h sidereal time +/- 1.5 hours at the date of discovery.  (Insufficient measurements are available right now to determine whether this node is associated with 0h sidereal time or with morning temperatures.)

3. Comparison with Lance Osadchey’s results suggests that peak-to-trough amplitude for measurements made at the same approximate sidereal time will vary with latitude.

4. The existence of this node demonstrates that the measured peak-to-trough amplitude is not dependent on floor slope so long as the floor is smooth and level to within typical household foundation tolerances (1 cm / 3 m).  Dr. Osadchey’s results are probably not due to normal variations in the slope of his desktop or a minor defect in his mount.

5. Measurement of thermal refraction in real time using a web cam reveals that while it can be of sufficient magnitude to explain the measured effect, it is too randomly variable to explain the observed effect.  Dr. Osadchey’s results are probably not due to thermal refraction of the laser beam while in transit between the laser and camera.

6. The computed standard deviation of the image center position (not the standard deviation of size) appears to be less than the anticipated value of 1, which implies a higher than computed statistical confidence.  (Note: In this case, statistical confidence only reflects the reliability that I am detecting something other than random noise.  It does not imply that I am measuring what I intended to.)  Without yet determining a causative agent, I can still say that, whatever I am detecting, it really is causing motion of the laser’s area of incidence relative to the camera CCD surface. 

7. Thermal drift is already apparent in the data, particularly for those data sets taken without air conditioning.  The dominant term does not appear variable. 

8. Thermal drift follows a cooling curve with a parabolic term.  This behavior can be derived from an equation for thermal bending displacement of a beam, and is observable in longer duration trials.  Given regular, angle dependent variations in the thermal gradient across the beam (such as heating a rotating beam using a fixed heat source), it is theoretically possible to set up a pattern of thermal drift which exactly mimics the anticipated pattern due to a hypothetical reference vector.  This false signal will vary from a real signal in two critical regards:

8.a. Its amplitude will be time dependent, not angle dependent.  The amplitude of the false signal will change with angle, but can be varied in proportion to the interval between measurements.  The amplitude of a real signal will not vary in proportion to the interval.

8.b. Buildup of thermal energy in the beam will limit the interval between measurements during which the resulting variations in thermal drift can mimic a real signal.  At some maximum interval, the thermal drift will wash out any real signal, smoothing out the curve of image motion. 

I also have a sufficient theoretical model of thermal drift to make a prediction regarding Lance Osadchey’s long term trial currently in progress.  I predict that he will find nothing.  There are three reasons for this.

1. Lance’s results posted to date demonstrate that two independently functioning Veladors of identical fixed orientation, operated in the same room, will return different results over time if not readjusted for drift before each measurement set.  Thermal drift, being strongly dependent on the structural properties of the support beam, can easily account for this.

2. Dr. Osadchey’s short term trials also show evidence of thermal drift.  His average observed drift appears nearly an order of magnitude less than my own, but I attribute this to more careful climate control.  Either observed drift (his or mine) remains sufficient to account for the observed motion in his long term trials over the course of twelve hours or more.

3. My mathematical model of thermal drift suggests that thermal drift will be cumulative over long periods, and will overwhelm any real signal of the measured amplitude.  (This is, in fact, one predicted means of differentiating thermal drift from a real signal.)

This does not reflect on the desirability of a long term trial.  However, a fixed orientation is not adequately sensitive to return useful results.  Over the course of an entire day, thermal drift will dwarf all other motion components.  A better method for a long term trial would be to conduct a daily trial with rotating orientations over a short duration and comparing the measured magnitudes over time, because this can isolate smaller magnitude effects.