Replication of “An Experiment to Measure the Absolute Motion of the Earth”

Lance Osadchey, C. Michael Edwards, and Pat Kirol

Dr. Lance Osadchey's article "An Experiment to Measure The Absolute Motion of the Earth" (The Citizen Scientist, 02 February 2007) was preceded by an introduction by SAS Founder and Director Shawn Carlson, who holds a degree in nuclear physics. Dr. Carlson wrote,

"This paper challenges one of the bedrocks of modern science, Einstein's Special Theory of Relativity. Lance Osadchey is a medical doctor who has carried out a series of experiments that seem to show that it may be possible to identify an absolute reference frame--a result that, if true, would cut out the beating heart of the Special Theory and throw modern physics on its ear.  

"Since I founded SAS back in 1994 we have received dozens of papers that challenged Einstein's ideas. I have reviewed them all, and I have passed several on to others to review as well. But we have never published any of them. Why not? Because they were all theoretical papers that were either grounded in postulates that violated the known laws of physics, or that contained significant mathematical mistakes. In short, they were clearly wrong on their face and publishing them would, in my view, not have advanced the cause of serious amateur science.  

"Dr. Osadchey's paper is different because he hasn't just sat in his chair and daydreamed about the subject. Rather, he has rolled up his sleeves and connected himself to the real world. Dr. Osadchey has carried out a series of experiments that have produced results that, while quite strange, contain a degree of self-consistency that demands an explanation. As someone who understands Einstein's theory very deeply and who also knows well the extraordinary level of experimental vindication that the theory enjoys, I would be astonished if it were ever overturned.  For that reason, I believe that we should all be highly skeptical of Dr. Osadchey's result at this time. Moreover, the paper does contains areas of concern that would probably prevent it from being published in a major peer-reviewed journal. Nevertheless, in my view Dr. Osadchey has done significant experimental work and has earned a fair hearing from his peers here in the citizen scientist community.  

"I hope that by publishing this paper we will stimulate the community to think, to experiment, and to debate. I look forward to your analysis, your insights and the results of replications of Dr. Osadchey's work."  

"Stimulate the community to think, to experiment, and to debate" is exactly what occurred, for Dr. Osadchey's article was followed by various letters in "Backscatter," a major article by C. Michael Edwards ("A Progress Report on Replicating An Experiment to Measure the Absolute Motion of the Earth," 6 July 2007) and the record number of views (6,733 as of this writing) for any topic on the Society for Amateur Scientists Forum.

All of this has also resulted in a carefully organized collaboration that will appear in two parts in The Citizen Scientist" and fulfill Shawn Carlson's call for "your analysis, your insights and the results of replications of Dr. Osadchey's work." Dr. Carlson's call for analysis, insights and results of replication remains in effect, and readers may send comments and results to The Citizen Scientist and join the discussion at the Society for Amateur Scientists Forum. Editor.

Introduction

We've been cooperating on the project “An Experiment to Measure the Absolute Motion of the Earth” (Lance Osadchey, The Citizen Scientist, 2 February 2007) for the past year, gathering data and working to meet a need for improved experimental controls. We've also received important contributions from other researchers.

Preliminary results were encouraging and appeared in the June 2007 issue of The Citizen Scientist (C. Michael Edwards, "A Progress Report on Replicating An Experiment to Measure the Absolute Motion of the Earth"). Our later results contradict the initial claim of an absolute vector reference and, therefore, do not challenge the special theory of relativity.

However, this does not in any way imply that the effect of interest has been accounted for. Our results also demonstrate that the magnitude and phase are both unrelated to heat transfer and independent of equipment specific parameters. This result contradicts some commonly proposed (but untested) explanations for similar observations by other researchers. Our results continue to suggest an orientation reference that may be constant on a geographic scale, and there remains a distinct possibility that we are observing an astronomical phenomenon.

We are continuing our joint experiment with the goal of identifying the cause of the observed deflection. We have also begun a second phase of the experiment in which we are testing models of a hypothetical external vector field as causal agent.

Instrumentation Changes

We are still using independent experimental plans, but we've derived a common procedure for conducting simultaneous control experiments. In addition to several temporary changes implemented as controls, one or both of us (Dr. Osadchey and Edwards) have made the following alterations to our instrumentation and procedures:

1. We doubled our minimum sample sizes (from n = 1 to n = 2 for Osadchey and from n = 2 to n = 4 for Edwards). This improved estimates of experimental error and better accounts for the drift inherent to an unstabilized laser source. n = 10 is employed for initial testing of new design variations (Edwards), but with damping time requirements this is too cumbersome for regular use.

2. We varied the elevation of our support beams to check vertical orientation responses.

3. We added thermal insulation and increased warm-up time (Edwards). Thermal mass of the laser cell was also reduced, with more clearance between the laser and mount. This reduced laser drift per turn to magnitudes less than the effect of interest.

4. Increased compression of the damping pads reduced measurement error due to pendular oscillations. (Edwards)

5. A magnetic compass was used to measure orientation instead of fixed position markers. (Edwards) This allowed a wheeled platform to vary position in the lab during trials without repeating the same position, floor slope and thermal environment for each turn.

6. An EMI filter circuit was added to the laser power supply and grounded shielding was added to the camera. This reduced image noise and improved accuracy (Edwards). The shielding is an electrically grounded transparent wire screen mesh with an aperture for the camera. Capacitors across the power terminals and in series with a common ground comprise an EMI filter circuit for the laser. And batteries are used to prevent constructive interference.

Figure 1. A schematic of the filter circuit for the laser power supply.

Switching out equipment for experimental controls revealed that some lasers already incorporate oscillator circuits. For these, the terminal bridge capacitor had to be omitted to avoid feedback.

7. We started an ongoing series of tests for geographic position dependence. This required transporting our instrument over long distances. A shorter beam arm was easier to transport (Osadchey), and more compact padded mounts have proven as accurate as other mount designs (Edwards).

Figure 2. The modified support beam stowed for transport.

8. We traded lasers by mail between our Vermont and Louisiana stations. This was the cheapest way to test equipment at both locations.

9. Analysis software was upgraded to automate the analysis (Edwards). We are still subtracting a polynomial fit for drift correction, but this fit can now be applied vs. time rather than angle, giving a more accurate approximation. A sinusoidal fit applied to the data gives more precise magnitude and phase values.

Horizontal and vertical axis data were analyzed independently. Magnitudes and phases were estimated from coefficients of a fitted sinusoidal function vs. angle.

where a is the measured instrument angle, a_x and a_y are the angles of peak magnitude for the x and y coordinates, b_x and b_y are offsets (typically 0.00 microns), and A_x and A_y are the amplitude components of the image motion. These particular empirical equations were derived based on a model of the deflection as due to a vector product interaction, but any sinusoidal functions of similar form can be used to obtain amplitude and phase relationships.

Experimental error was estimated from the standard deviation of the actual data vs. this fitted curve.

where X_j, Y_j and a_j are the computed mean coordinates and instrument angle for increment j of m increments per turn over n turns. Trials with s > A were rejected as unusable; however, s_x and s_y are assumed independent.

10. Two laser-camera systems with opposed direction on the same support beam are now used as a reference (Edwards). Both camera triggers are wired to a single switch. This arrangement highlights coupled motions, and provides simultaneous control data for phase and amplitude comparisons.

11. We plan a set of control experiments utilizing a reflective mirror to check for geometric relationships inherent to some models of light aberration. These measurements will serve as a follow-up to experiments with that geometry performed by Pat Kirol, which were inconclusive.

Signal Isolation

We identified several causes of systematic error in addition to those investigated previously. Each was investigated for its potential to produce a false signal.

1. Inherent Image Noise – In addition to behaviors identified previously, we investigated means of controlling feedback in the camera circuitry, which can cause scrolling distortion and saturation of the recorded image.

2. Laser Drift – We have identified laser drift as the dominant source of cumulative error in our experiment. Variations in the laser's power curve and load slightly alter its emitted wavelength, and heating of the laser diode alters its surface dimensions. Both phenomena change the laser's interference pattern, causing motion of the central peak without changing the direction of propagation.

3. Electromagnetic interference – EMI from wall current induces a variable component in the power curves (even with battery power), the phase of which varies with the orientation of the instrument. This component causes constructive interference in circuitry powered by regulated power supplies. This in turn can create drift variations and feedback due to fluctuations in laser output.

These effects, as well as those previously identified, were all observed to some extent.

Inherent image noise can be identified by visual inspection of the photographic data. Typical patterns are shown in Figs. 3a to 3f.

Figure 3a. Overloading of CCD at central peak.

Figure 3b. Internal interference in the CCD photodiode (closely spaced fringes below the peak).

Figure 3c. A linear structural defect in the CCD photodiode.

Figure 3d. A contaminant on the photodiode.

Figure 3e. Scrolling line distortions visible across the image.

Figure 3f. EMI induced saturation varying with orientation.

The magnitude of image noise is instrument specific and varies with the camera used.

Overloading of the CCD (Fig. 3a) overwhelms the cells at the image peak center. This can be compensated by manually filling the center with the color of the bordering cells to smooth it. Overwhelmed cells are otherwise neglected in analysis.

Interference fringes caused by internal reflection in the photodiode (Fig. 3b) are typically only a few pixels in width, but make the peak boundary grainy. This creates measurement errors, but won't oscillate without a change in intensity or angle of incidence. Structural defects in the photodiode (Fig. 3c) and shading by contaminants (Fig. 3d) also create errors but do not oscillate.

Figures 3e and 3f show scrolling distortion and saturation. Scrolling manifests as moving localized expansion or contraction of linear sections of the image. These distortions are typically aligned along either the vertical or horizontal axis. Saturation (e.g., fluctuations faster than the CCD can register) obscures the laser interference pattern, making it impossible to isolate the central peak and creating an asymmetric expansion and contraction of the laser image. These can be eliminated by smoothing the laser power supply curve to reduce feedback in the camera.

Although all image noise obscures the effect of interest, scrolling and saturation can mimic a real signal. The period of scrolling is independent of the instrument's period of rotation, so scrolling will not generate a repeatable false signal without a resonance between the scrolling and rotation frequencies. Saturation is sensitive to wall current EMI, and varies regularly with orientation. However, it is instrument and location specific. EMI induced saturation can be controlled for using simultaneous measurements with multiple instruments, changing instrumentation between measurements, changing filter circuit settings and using multiple locations.

EMI can potentially explain phase correlations over geographic distances. Most power plants in North America (where all of our observations to date were made) run their generators in phase to allow transfers between power grids without destructive interference. Thus, most wall current in North America is in phase. These power plants use the atomic clocks of the GPS system as a reference to maintain an average drift rate of less than 3 Hz per year. EMI with a daily phase variation of more than 3º would require a regulating mechanism (e.g., temperature induced inductance changes). Such a regulating mechanism would be equipment or location specific.

Laser drift was observed in all of Edwards' trials and in long duration trials conducted by Kirol. It will change rate and occasionally reverse during a trial. Control experiments with rotation of the laser between turns show variation of the laser drift, but do not show inversion of the effect of interest. Adding insulation and heat sources also affect laser drift, but do not change the magnitude or phase of the orientation dependent deflection we're seeing. This implies that the effect of interest is independent of both cumulative laser drift and external heat flux. The oscillations are not driven by an external heat source.

Stabilization of the laser at the detector (a common method of drift compensation for large interferometers) was rejected because it edits out the effect of interest. Laser drift is reduced by allowing long warm-up times for the laser, increasing thermal insulation around the laser mount, reducing laser power supply voltage, and by judicious choice of laser. Laser drift is time dependent, and it can be minimized by decreasing the time interval between photographs. We frequently reduce laser drift to a rate for which the displacement due to laser drift over each turn is smaller than the effect of interest. Thus, variations in cumulative laser drift may be eliminated as a cause.

Figure 4. A graph of horizontal image position for two turns with external loading of the support beam, one clockwise and the other counterclockwise, illustrating hysteresis due to sagging under load.

By alternating the direction of rotation with the motive force applied directly to the instrument support beam instead of the mount, we identified a hysteresis effect due to sagging of the support beam in the direction of the motive force (see Fig. 4). This variation depends on the direction of applied force, not instrument rotation. It does not appear in instrument designs that don't transfer torques to the support beam.

In principle, use of a reference with light propagation opposed to that of the test beam can be used to differentiate between geometric behavior of different models of this deflection. In the lab frame, thermal responses should produce deflection in consistently aligned directions for both the test and reference lasers, with an amplitude ratio randomly determined by the instrument material properties. Classical light aberration due to linear acceleration would share this geometric relationship. Deflections due to an external vector field with a vector product relationship can occur in opposite directions in the lab frame, with roughly identical amplitudes.

Figure 5. Scenarios for: laser drift due to a heat flux (Q), deflection by an external field (A) and support beam bending under coupled moments (M).

Figure 6a. Expected response of lasers with opposed directions in a horizontal flow field.

Figure 6b. Expected response of lasers with opposed directions in an inclined flow field.

For example, Figs. 6a and 6b show expected behavior for light aberration due to a hypothetical flow field or linear acceleration. The deflection is always in the direction of the flow field. The local coordinate planes of cameras A and B are shown. B is rotated 180º so that its x-axis is in the opposite direction from A, so the measured x-coordinate is of opposite sign from A to B. The apparent reversal between A and B does not occur without this rotation. No rotation = no reversal.

In the majority of our experimental trials, the instrument is rotated in the horizontal plane only. The y-axes of A and B are in the same direction and remain so for the entire trial, with no vertical rotation of the instrument. Thus, a reversal of vertical deflection between cameras A and B would show that the causal agent is not aberration due to a flow field.

Similarly, light aberration due to a vector product interaction should produce a deflection nearly opposite that of propagation in the opposite direction in the instrument frame. It is possible to test for this geometric arrangement by reflecting the laser from a flat mirror, folding the light path nearly 180°. The predicted point of incidence on the photodiode will vary between scenarios. Classical light aberration (e.g., a flow field) will show an exaggerated deflection, as the increased angle of incidence changes the angle of reflection. Deflection would then be increased by repeated light aberration on the return transit. Conversely, a vector product interaction could counteract the change in angle of incidence per Fig. 7, producing no change or a net decrease in magnitude.

Figure 7. Scenarios for near 180° reflection in three causal field models.

Readers are encouraged to send comments about this article to "Backscatter" at editor[at]sas[dot]org. Part 2 will appear in the August 2008 The Citizen Scientist.