Reproducing the Velador Experiment

During preliminary calibration, I performed an initial measurement trial on 4/23/07, using a CFI resolution webcam. 

This calibration run revealed important shortfalls in my current design.  The resolution of my initial setup, with CFI resolution vs. the 4 Megapixel camera used by Lance Osadchey in his reported experiment, is 16 times less than Dr. Osadchey's.  Mathematically, this yields an estimated pixel size of four times larger (i.e., my velador can only measure motions larger than four times what Dr. Osadchey's can).  The derivation of this is discussed under camera selection, and was anticipated.  However, I failed to account for the fact that each CCD cell requires some minimum light level to be triggered.  In a uniformly varying image, this means that the actual effective size of each pixel is six times larger using my camera, because the fringes of the image may not necessarily have enough light to trigger the CCD unless they extend completely across the CCD cell.

The image from Dr. Osadchey's experiment discussed in the page on measured thermal distortions shows that the expected range of motion is on the order of six pixels using a 4 Megapixel camera.  A CFI camera does not have sufficient resolution to confirm the effect I'm looking for.  I will require a new camera with improved resolution.

Also, my initial calibration measurements confirm that Dr. Osadchey's setup, with use of an attenuating filter and removal of the camera lens to allow the attenuated laser to shine directly on the CCD, is the superior setup.  In an effort to save money and develop an experimental design suited to a camera whose lens cannot be readily removed, I solved my attenuation problem by interposing a translucent membrane instead of an optical filter.  Paper and tinted polyethylene were both used, with polyethylene yielding the best results.  However, internal reflection in the translucent material tended to distort the edges of the image, and concealed the surrounding interference pattern.

The image motion is too small to be visible from these four images.

This image was prepared by creating a false color image of the East and West orientations (a 180 degrees turn of the apparatus) and superimposing them.  MS Paint was used to prepare the images, and the same procedure was used for both: conversion of *.jpg images to 16 color *.bmp format, followed by selected color changes, then superposition of the adjusted images.  Care was taken to make the overlap exact.  (All four image corners should match exactly.)  The focus was on the central portion of the image, because diffusion of light in the screen increases with distance from that boundary.  The blue pixels are from the central region of the East Image, while the green pixels are from the central region of the West image. 

Please note that the majority of the blue pixels are found along one side of the superimposed real image.  The concentration of blue pixels is more uniform for the ghost image.  This implies motion of the real image across the screen between the two frames.  The average difference in position between the two photos is approximately 1 pixel across the entire central area of the image.

Given that this motion would be 4 to 6 times as many pixels using a 4 Megapixel camera, this strongly indicates that I am observing the same degree of motion as was reported by Dr. Lance Osadchey in his March article in The Citizen Scientist. 

I suspect I am seeing the same effect, with the same approximate magnitude.

This is not a successful outcome yet.  The observed motion, while clearly of statistical significance, is still not  greater than the level of noise (combined thermal and artifacts of the screen design) that is also present in these measurements, and is at the upper limit of measurement error.  These images must be cleaned up in order to confirm Dr. Osadchey's results.  A broken line of pixels is not sufficient proof.  That means a new camera and a new filter design.  This will require more money (up to another $200), and that will delay construction for up to two weeks or more.

However, the fact that this initial result is consistent with Dr Osadchey's reported result is very encouraging, and I fully intend to continue this experiment.  I will be conducting a few more measurements with my current setup, primarily for calibration purposes, and will begin work on an improved mount for the support beam to allow measurements at other elevations than horizontal.  Most of the reconstruction related to the camera can be confined to the camera and laser cells, which are removable without a complete rebuild.