READERS OF THIS department will recall a discussion of air currents in reflector tubes, by F. N. Hibbard, of Richmond, Virginia, in the January, 1944 number. The following discussion on this subject is by E. K. White, Chapman Camp, British Columbia, and is printed from The Journal of the Royal Society of Canada, Toronto, Ontario.

"PRESENTED in this article are the results obtained from experiments made with an electric fan forcing a current of air into the metal tube of a Newtonian reflector.

"The telescope is of 9" aperture, and 100" focal length, and is used chiefly for the study of lunar and planetary detail. The tube is of iron, 10" in diameter and 104" long. There are no ventilation holes in the tube, save for a narrow space between the wall of the lower end of the tube and the mirror cell. When not in use the instrument is housed in a roll-off wooden shelter.

"It is well known that many observers are of the opinion that closed metal, and even closed wooden, tubes have a detrimental effect on the seeing, the reason being assigned to the presence of air currents within the tube of a different temperature from that of the outside surrounding air. Possibly the best remedy is an open skeleton-type tube, similar to those of the large reflectors. However, one living in a damp climate, as prevails in most of southern British Columbia, will have severe dewing and frosting conditions to contend with unless the mirror is removed to the house when not in use, and this method is bad for the silver film. The closed tube insures freedom from dewing of the mirror and prevents trouble from stray light; also, it aids in keeping the mirror free of frost crystals in winter.

"The late W. H. Pickering, ('Amateur Telescope Making-Advanced,' pages 610-612), well known lunar and planetary observer, claimed greatly improved seeing with closed tube reflectors by forcing a stream of air through the tube, introducing it by means of a small fan near the mirror's surface, and expelling it out of the open end. Here is movement of air within the tube, but of the same temperature as that outside. In other words, the troublesome warmer or colder air in the tube is displaced by outside air. Of course, the fan must be kept running as long as observations are made.

"By way of experiment, a round hole 5" in diameter was cut in the lower wall of the tube, just above the surface the mirror. This is covered by a metal hinged door, having a close-fitting felt gasket. A 5" electric fan was obtained and experiments started.

"At the beginning of observations the extra-focal rings of a bright star, usually Polaris, were noted. Immediately the tube door was opened these rings became much distorted and unsteady. When the fan was started the rings became circular again and steadier, but they always showed a rapid rotation in the direction of the fan blades. With the fan running, the rings were definitely steadier than when the fan was off but the door open.

"The definition of detail on the moon and the planets Saturn and Jupiter was studied, with and without the fan, in all conditions of seeing the skies afforded. It was impossible to note any improvement with the fan on these objects, when the seeing was between 0-4 (scale of seeing objects devised by W. H. Pickering very poor, 10 excellent). However, on a few occasions when seeing was estimated to be from 5-7, a very slight improvement in definition seemed certain when using the fan. The floor of the lunar crater Plato under oblique lighting was used as a test, and on good nights from two to four craterlets could be seen as such. With the fan running, these marks were steadier although no more fine details could be seen than before. No further detail could be seen on the planets mentioned, simply a steadying of the difficult marks seen without the fan. No improvement could be noted in the resolving of close double stars in any condition of seeing, but little time was spent with this particular investigation.

"The fan must be mounted on an adjustable stand, and not attached to the tube, so as not to set up any vibrations in the instrument. The blades should be kept close to the tube opening, in order to have a maximum current of air flowing through it. These requirements introduced some inconvenience in any careful observing program by necessitating rather frequent adjustments of the fan as the drive carries the long tube forward.

"From results so far obtained, with the exception of rare nights when seeing is quite good, the use of a fan is not considered worth while as a measure to improve definition. While it is known that other observers than W. H. Pickering have claimed that the use of a fan gives marked improvement in definition, I have unfortunately been able to find little published of their findings, or methods. Further experiments with a fan are being continued at this station."

THE FOLLOWING is the remaining part of an article begun last month by Daniel E. McGuire, on the use and technique of scotch-tape spacers for speeding up the previously tedious task of testing optical surfaces by means of interference fringes:

The collimating lens system can be computed by graphic ray tracing. This is done for a zone about 90 percent of the way from center to edge of the test plate surface. This zone focuses to the center of the smallest circle of confusion in the image formed at the viewing distance. The single surface collimator is easiest to compute by graphic ray tracing (Figure 3). It is much more difficult to determine the amount of curvature required for the two lens collimating system, especially when both curved surfaces are required to have equal radii in order to simplify tool making.

The designing may be simplified by selecting a curvature for the back of the test plate which deviates the ray through approximately one third of the angle required to focus to the eye position. The Hat surface of the auxiliary lens further deviates this ray. The last step is that of finding the radius required in the third surface to deviate the ray still farther, so that it crosses the optical axis at the required viewing distance. In the first and third steps, an unknown curvature is determined by means of a known index of glass and a known deviation. In the second step, an unknown deviation is determined by means of a known index and curvature (or flatness) of the glass. Only the first step is used with the simpler system.

The two diagrams in Figure 3 show how to determine the radius of the single surface collimator. The test plate curvature is indicated by R₁. The clear aperture falls within the dotted line, and the 90 percent zone is used in the calculation. The radial line passing through the 90 percent zone locates point P laterally, but its longitudinal position is judged according to the required thickness of the glass at this zone. PE crosses the axis at the eye position E. Points A and B are distant from P in the same ratio as the refractive index of glass and of air. The index in this case is 1.52, so PA equals 1.52 times PB. Line CD is made parallel to AB, passing through point P. The center of curvature of the collimating surface, R₂, is found where line CD crosses the optical axis.

The concave test plate (Figure 3, left) nearly exceeds the limit of curvature for a single surface collimating system. Spherical aberration is much reduced by using an auxiliary lens.

The convex auxiliary lens must be made larger in diameter in order to include the marginal rays. In Figure 4, the auxiliary lens is too small to illuminate the margin of the test plate; but it is advisable to make the test plate oversized, with illumination over the needed area. In this way, a turned down edge on the test plate can do no harm; it is masked off, invisible.

The convex test plate is well within the limit of practical design. When it is necessary to add an auxiliary lens, it is made still smaller in clear aperture at its back surface to keep its edge thickness to a minimum. Its plane surface rests upon the flat area surrounding the first collimating surface, with the concave side toward the eye. The same rule can be applied, regarding the use of an oversized test plate, to exclude from vision an imperfect margin of the surface.

The 100 percent zone is at the edge of the required diameter of test plate surface. An additional 5 percent can be added to the diameter of the polished surface to simplify the making of a perfect surface within the used area. Ray tracing the 100 percent zone shows the required boundaries of collimating lens surfaces.

For a further study of graphic ray tracing through a number of surfaces see Martin's "Applied Optics," Vol. I, Figure 7.

When designing collimating test plates, it is not desirable to have normal incidence for the marginal rays on any of the collimating surfaces. An optical flat, having a plane-parallel back surface, needs an auxiliary lens to collimate the light, but the light reflected from the back surface is focused at the eye position the same as the light used in testing. This fogging of the fringe pattern can be reduced by a non-reflective coating on the back surface of the test plate. Non-reflective coatings on all collimating surfaces improve the test when extraneous light from the room is reflected in the surfaces. Reflection of the light source in the center of uncoated collimating lenses is not objectionable.

There are two methods of using the test plates. Figure 1 shows how finished work, or single surface work in process of polishing, may be tested. It is practical only where the work is easily handled by the edges, or by an attachment. The work is viewed in the mirror below. Nothing is reversed in the mirror, since the work is interpreted as though looking down from above. Top, bottom, left, and right in their respective places. This method eliminates handling the test plate for each test. Only the work is touched and for a very short period per piece and thus thermal effects are negligible. Some 300 to 400 surfaces an hour may be tested.

The test, as shown in Figure 2, requires handling of the test plates but cells of heat-insulating material are used to guard against thermal effects. Thin edged, convex lenses are difficult to test with the arrangement in Figure 1. In Figure 2 they are placed on a padded ring, with the test plate on top, while looking down through the beam splitter.

With sufficient clearance between beam splitter and table, multiple blocks of lenses are tested, as shown in Figure 5. The lenses in the outer row are viewed in turn by rotating the block upon a fixed axis that is set at the proper angle. The block is set with axis coincident with that of the light for testing the center lens, or at a slight angle for an inner row of lenses.

In order to simplify the drawing, Figure 5, the tool shank is made longer than it ordinarily is. A shorter shank requires a convex block holder to fit roughly into the hollow, curved back of the blocking tool. Tapered holes are bored at the proper places, similar to those shown in the flat sided block holder.

A concave block of lenses requires a concave block holder, or simply a hollow arch, with holes bored at the proper places to line up the various rows of lenses with the axis of the light source. Blocks of flat work are held in place by any one of a number of adapters all of which are mounted on a flat plate. Small, light-weight work is easily supported in one hand while making the test.

Air-spaced test plates are not entirely new, although the writer may have some original ideas. These, or similar, methods are being used in some optical shops-just how many and in what circumstances is not known.

The Society for Amateur Scientists (SAS) is a nonprofit research and educational organization dedicated to helping people enrich their lives by following their passion to take part in scientific adventures of all kinds.