TWO MONTHS AGO in this space we described Broadhead's mirror grinding technique, last month the question of thin mirrors was discussed, and now the two are conjoined: Broadhead making a pair of thin mirrors.

Recently, Dr. John Strong, author of the widely used book "Procedures in Experimental Physics" and now professor of experimental physics at the Johns Hopkins University, asked this department whether some of the advanced amateur telescope makers could make, for non-astronomical research in the infra-red, a matched pair of 36" spherical mirrors of 1-1/4" plate glass (thickness-to-diameter ratio, 1:27), of focal length between 99' and 101' (f/33) and in focal length not more than 6" different from each other, with a circle of confusion (pinhole image) at center curvature not greater than 1/8" diameter in single reflections. These mirrors were to be centrally perforated with holes, mounted in metal cells, and delivered within 90 days. Thereafter, at the Johns Hopkins University, they were to be set in an air-tight tube 100' long, which tube could be partly exhausted and cooled to simulate all atmospheric conditions as high as and including the stratosphere. The set-up is like that shown diagrammatically in Figure 1. (Without the small mirrors to offset it the light beam would inescapably return to the pinhole, no matter how the mirrors were tilted.)

This job went to Dave Broadhead, Wellsville, N. Y., whose precision optical background was as follows: Prewar, had made a small objective lens, 6" and an 8" reflecting telescope, and in wartime 3000 roof prisms with better than 98 percent acceptance rate; postwar, two 10" paraboloids, three 14" flats, about 6/7 of a Maksutov telescope (now nearly finished, and so are several other "Maks"). How he went about doing the two 36" mirrors, what methods he used, should interest all fellow amateurs, particularly those who secretly harbor hopes of someday making larger than average mirror, which means nearly everybody. The account is especially valuable because Broadhead does not try to hide or minimize his mistakes. He delivered the mirrors, better than specifications called for and a month sooner than they called for, and these earned Professor Strong's comment and OK to quote: "We are just plain pleased, from the ground up."

Offhand, these looked like dead-easy mirrors. Broadhead's cellar shop, 12' x 36', the neatest, brightest, and cleanest we have seen, affords a lathe drill press, metal shaper, grinder, electric welder, gas crucible furnace, and room. Most of these proved to be essential. The less than full precision tolerance did not please but disappointed Broadhead, as it would any true amateur, but there were offsetting headaches and uncertainties, also one certainty- the 90-day delivery limit. Hence, all in all, the job contained enough adventure to interest. Here are some excerpts from Broadhead's progress reports:

"Have to build a special machine (Figure 2). Vertical spindle of 2" shafting, on top of which is a round mirror support welded up from 1/2" boiler plate, with radial ribs and braces (Figure 3). Very rigid 24" grinding tool (Figure 4,) faced with glass blocks as in Strong's book page 44, rotated by a 1/8 h.p. motor. Main oscillating arm of two-by-fours driven by a separate 1/3 h.p. motor. Tool on top. I have an ammeter in the driving motor circuit as a guide and warning to any approaching sticking of grinding tool, since a stuck 24" tool wouldn't be funny.

"Polisher (Figure 4) 31" diameter, made of 1/2" boiler plate, radial fins welded on its back side. Face has 3/4" pitch layer, 2" facets. Takes 150 pound pull to budge it. Weighs 180 pounds.

"For knife-edge test on the 200' course I located image of a 1/64" pinhole with a big piece of cardboard. Found I had hit the 200' radius within specified limits first whack, using a 22" spherometer I made. (Since I didn't have a 22" flat with which to set its zero point I set it at zero on the concave mirror and then on the fitting convex tool. The difference, divided by 2, gives the desired sagitta and that with double accuracy.) But there isn't much to see at the knife-edge. Usually the convection currents make the image look like a three-alarm fire (Figure 6, reproduced at exact size). One night, just before a rain, the air was steady enough for me to see the figure. However, I don't try to use the knife-edge a figuring guide. Instead, I use an 8" test plate and study the interference fringes between it and the mirrors. I use the shape of the circle of confusion merely as a rough indication of astigmatism (Figure 6, left). Incidentally, I may bring on broken arches from lugging these 140-pound mirrors upstairs and outdoors to test them, and neuresthenia from lugging them in again for fear of dropping them.

"While grinding the 36" I gave the 8" test plate one go on it for each grade of abrasive, then polished the mirror a bit to see whether the spherometer was giving me the right answers. Then I polished the test plate, trusting to experience to get a sphere. So, actually I got the test plate right by using the spherometer in a roundabout way. Of course, it would be practically impossible to test an 8" test plate at 200' radius; at least much more difficult than the way I took.

"At present No. 1 mirror has a maximum variation from spherical of 3/4 wave in any 8" area. No. 2 is ground through 400 Carbo. Had it polishing once but got a bad scratch. Mirror so light (except when you are lugging it around) that it is lifted by the grinding tool by suction when tool is lifted with a chain fall. After a lot of teasing it suddenly lets go, the mirror clumps down 1/2" or so, and the tool seesaws and thumps the mirror. Heart failure every wet."

THREE WEEKS later, job finished, Broadhead completes his report, and in what follows he very frankly points out the main mistake he made, and "don't we all?" (If the log book for the 200" mirror is ever published in toto, would you be critical if mistakes that must have been made by the human beings who made it were similarly set forth with candor, or would you expect the makers of that mirror to be different from other humans? If so, why?

"Radius of mirror A is 198'3", that of B is 199'1"; Circle of confusion obtained by 2/3 diameter grinder and 5/6 diameter polisher and proper strokes came well within 1/8" tolerance specified for single reflections without recourse to local figuring. One modification of my previous ideas and technique that I learned-though it should have been obvious-is that my rigid tool technique is useful only up to the point where the tool is as stiff as the work. I erred by making the grinding tool probably several times as stiff as the glass, and thus unnecessarily heavy, and this in turn bent the glass and gave greater astigmatism troubles. Yet, to avoid astigmatism, I was trying to support the glass perfectly. I could have arrived at a better balance of factors. If I had it to do over again I could, if it served any purpose, make the mirrors much more easily, using the original 2/3 tool for roughing, a 5/8 tool of aluminum for fining, and a similarly light lap backing for polishing (in order that, if it became necessary to return to grinding, the glass blocks need not be re-cemented on the tool).

"Although the mirrors are well within all specifications, it is instinctive for a TN to want to continue till his best level of workmanship is reached, also until every shred of information has been wrung out of the job. In this sense, such a job as this was disappointing.

"I found that the most convenient bed for the mirrors was the long-used Brussels carpet, but that the support given by it varied considerably with the direction of the nap. Here rotation of the mirrors at the end of each wet helped, but on the rougher grades, say 180, the variation in spherometer readings on diameters at right angles, due to the lie of the nap, was about 0.0005" in a single wet. At 600 emery this was reduced to 0.0001" and after several wets of UO10 Garnet Fines I could no longer detect this astigmatism with a spherometer. Yet, after a short polish, I could easily detect plenty of it with the 8" test plate. It slowly reduced during work with the 5/6 lap, but still was much in evidence even after the polish was complete and the circle of confusion reduced to about half the tolerance specified. The advantage of the Brussels carpet is that the variation in the support is fairly constant, whereas with a hard support the mirror touches at only three wandering places.

"Had this been an astronomical job the outstanding astigmatism would have been ratable as heavy, and would have called for slower methods and thicker blanks. The thin glass was used simply to save costs in laboratory research with these mirrors that did not call for a more expensive set-up."

COMMENTING on this job, Professor Strong stated: "The difficulty of making these thin mirrors was an unknown. I even had to make some computations before I was satisfied that such a mirror would hold a figure once it was given one; and I found that it wouldn't hold a figure unless it was used in a fixed, vertical position. Very thin mirrors can be used in a fixed vertical position if they can be given a figure. That was my joker: Who could give such thin glass a figure? I was convinced that it could be figured by someone with imagination, skill, and enterprise, but that conviction was based more on faith than on facts. Now we have the facts, and 30 days ahead of schedule. The circle of confusion for the final image, after four reflections, two from each mirror, was less than 1/4" in diameter. The light in the final image of a Western Union point source after a 600' travel and four reflections from Broadhead's two mirrors, was a hard core of about 1/8" diameter, and practically all the light fell within a 5/16" circle. For our needs this is perfect performance."

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.