Reproducing the Velador Experiment
Expected Manifestation of Thermal Refraction
(Last reviewed 6/5/07)
Given an axial thermal gradient equal to or greater than the radial thermal gradient (the configuration creating the least spread of the laser) the estimated minimum temperature difference across the interface necessary to create a deflection of 30μm over a 3 m beam path is estimated using equation 9, with the assumptions shown. The computed value of thermal gradient across an effectively 2 dimensional interface is 9.5 K.
This gradient far exceeds any expected value, indicating that the effect of axial thermal gradients is negligible if the apparatus is not set up alongside the equivalent of a household furnace. Radial thermal gradients will be responsible for any thermal refraction of the laser beam large enough to interfere with measurements, and radial thermal gradients will create noticeable beam area spreading as a side effect (up to 100% or more of the observed edge motion).
Radial thermal gradients are only of interest at the support beam axis – the region through which the laser beam travels. Radial gradients at the axis can be reduced by introducing fan-forced flow through the beam. A wire mesh screen across the support beam cross section would also be necessary to induce laminar flow in this case. This could introduce complications with aligning the beam path, and is not recommended for the first experimental trial.
Visible thermal distortion of the image can be created using a significant source of heat, in order to investigate the effects of a strong radial gradient. Directing the beam through the edge of the updraft over a candle flame provides a very steep thermal gradient of several dozen degrees Celsius, well in excess of the estimated gradient for 30μm distortion with an axial thermal gradient. It also provides a small surface on which the radial thermal gradient is sufficient for internal reflection.
Image Without Thermal Distortion
The laser was positioned approximately 3m from the wall it was projected onto, with a lit candle approximately 30cm in front of the laser. No mount was used for the camera, and the images were taken with the camera hand-held. The laser was passed through the candle updraft at a position approximately 1cm above the visible flame, and slightly within the visible radius. The resulting image distortion was photographed.
.
This photo is of the undistorted beam incident on a paneled wall. The shadow is the photographer’s fingers, the whiter image is the actual beam image, and the other darker image is a ghost image due to internal reflection within the camera. A panel groove is visible on the right side of the image. The width of the white portion of the real image is approximately 2mm wide and 4 mm high. The vertical axis is aligned with the panel groove.
Image with Thermal Distortion
This photo is of the distorted beam on the wall. Note the slightly different position of the shadow and ghost image relative to the real image. The real image is significantly distorted horizontally. The beam’s incident area has been stretched out horizontally by passage through the relatively extreme thermal gradient above the flame. It is not discernable from the two pictures because of the slight difference in image scale and camera position that results from photographing without a camera mount, but the laser beam has also been refracted so that the outer edge of its incident area is moved 2mm toward the panel groove. The beam area has been stretched in the direction of refraction.
Even with the relatively enormous local gradient involved (hopefully more than one hundred times what the assembled Velador will see during operation, although no attempt was made to measure it for this trial), the beam distortion is still not sufficient to move the incident area more than a few times its minor diameter. However, it can be observed that distortion of the incident area occurs on the same approximate scale as its observed motion.
Checking for Thermal Distortion in Velador Image Data
It is possible to examine Osadchey’s data for similar thermal distortions. Photographs of two beam positions at relatively large angles from each other are presented below. (These are images 001 and 008 of data posted to the Citizen Scientist, March 2, 2007.)
These images were taken with a fixed camera mount as well as a fixed laser mount, which means that the spatial relationships between identical positions on each image do not change. Using the animation of this image set available at the Society for Amateur Scientists’ web site, several small sections of each image that do not change regardless of the observed motion of the laser beam can be identified. These sections correspond to lint or similar distortions on the CCD chip surface, and can be used as markers for measurement. The horizontal positions of four of them (Labelled A, B, C and D) are marked in the image above, and their positions do not change from image to image to within a resolution of one pixel. Lavender lines (labeled 1 through 6) mark characteristic boundaries on the incident area of the laser beam and its diffraction pattern. These points move with the beam, and should show a proportionate distortion if the motion is due to a thermal gradient with a component perpendicular to the path of the laser beam.
Just as there is no relative motion between lines A through D, the observed relative motion between lines 1 through 6 is 1 pixel or less, with a mode of zero pixels, despite the fact that average horizontal motion of the image between frames is measured at 3.7 pixels. The average difference in separations is only 0.3 pixels, in comparison to a measurement error of 1 pixel for this comparison.
There is no confirmable spreading of the image area. No significant thermal refraction of the laser beam path is occurring in these two images.