Grantville Gazette, Volume 71

by Bjorn Hasseler

  So there is an incentive to at least start thinking about military searchlights . . . . And conceivably small searchlights would be advantageous for nighttime civilian use, too: spotting navigational hazards, rescuing men from the water or a disabled craft, and signaling.

  There is a strong kinship between ship searchlights and lighthouse lights. Of course, the latter can be much larger and heavier.

  Light Sources. Electric searchlights, with light generated by a carbon arc, were used at the siege of Paris (1870-1). In a carbon arc, a strong electric current is made to flow across a short air gap between two carbon electrodes. The proof of concept was made by Davy in the early nineteenth century. Grantville literature provides some design guidance (EB11/Lighting, 659-66).

  The arc can be started only by bringing them in contact with each other, but then the electrodes are slowly separated. Since the rods burn away you need a mechanism to maintain the arc gap. The stability of the arc is improved by putting a ballasting resistance in series with it (which increases the power requirement).

  Direct current is preferred as it causes the anode to form a crater, which gives off most of the light. The intensity is greatest at a 30-45o angle from the anode axis, and this facilitates capturing the light with the reflector (Baird). High currents (130-300 amperes in 1917) are used in military searchlights so the source must be close by.

  To provide the direct current, the carbon arc light would be powered by a dynamo (a type of generator). The first dynamo was built in 1832 but major industrial use (e.g., in carbon arc furnaces) didn't come until after improved designs were patented in 1866-7. Electrical engineers in Grantville would know how to design a good dynamo.

  In NTL, carbon arc lamps are in use in Grantville in October 1633; see Offord, "A Season of Change" (Grantville Gazette 50), and at Rasenmühle in April 1634, see Prem, "Ein Feste Burg, Episode 7" (Grantville Gazette 46),

  The first carbon arc lamp emitted over 10,000 lumens (Banke), and I found an ad for a 60-inch WW II carbon arc searchlight that put out 525,000 lumens (candlepowerforums). Carbon arc lamps have low luminous efficacy (2-7 lumens/watt) and efficiency (0.3-1%). Hence, they generate a lot of heat; consideration must be given to providing proper ventilation.

  Now, it is worth noting the power requirement for a searchlight-scale carbon arc lamp. The US Navy Model 24-G-20 24-inch searchlight used in WW II was operated at an arc current of 75-80 amperes and an arc voltage of 65 to 70 volts. However, the line voltage was 105-125 volts, so almost half the power was absorbed by the rheostat/ballast (General Electric). That corresponds to a power draw of 7875-10,000 watts. If we assume 80% efficiency in the generator and distribution system again, then we would need as much as 12,500 watts, and thus a steam engine of about 17 hp. That seems doable.

  In fact, the Royal Anne, an airship built in Copenhagen and first flying in September 1636, has six steam engines (Evans, "No Ship for Tranquebar Part Two" Grantville Gazette 28), and I suspect that these steam engines correspond to those that Evans proposed for a medium-sized cargo airship in his "Wingless Wonders" (Grantville Gazette 19). Those engines were nine-cylinder, single-acting, "with 300 hp generated when running at full speed (2200 rpm, 400 psi)."

  ****

  If electricity is unavailable, there is a chemical alternative. Limelights, invented by an ordnance survey officer in 1822, were first used theatrically in 1836. They relied on the reaction of oxygen and hydrogen gases with quicklime (calcium oxide). That reaction is potentially explosive, and the safest format is one in which the two gas jets meet at an angle where the lime cylinder is located (Encyclopaedic Dictionary of Photography 303).

  Limelights were used by the Union Navy during its bombardment of Charleston in September, 1863 and to spot blockade runners in early 1865 (IATSE, KCWB, Navy 1). Drummond used the lime light (supposedly equivalent to "about 265 flames of an ordinary Argand lamp used with the best Sperm Whale oil") in conjunction with a 21-inch parabolic reflector for geodetic purposes; the combination produced about 92,000 candlepower. While he urged its use in lighthouses, the American Lighthouse Board reported in 1868: "The Lime light required much labor, there was danger associated with the production of the gases used, it required expensive apparatus, and the liability of the lime to become deranged far outweighed any advantages in the way of superior illumination, which could be derived from it." (USLS).

  Some sort of chemical-based searchlight was still available for military use in the early twentieth century, but its useful range was something like one-eighth that of the 36-inch electric search light (Ordnance, 37).

  The navy would likely rather use carbon arc searchlights, on both safety and performance grounds.

  ****

  Light Concentration. Note that the "candlepower" (light intensity in the direction of the target) of a light increases if its light is more tightly focused, even though the total light output is constant. A searchlight may have millions of candlepower in its beam. Light may be concentrated by mirrors, lenses, or combinations of the two.

  Reflectors. The earliest documented use of a polished metal reflector to concentrate candlelight was in 1532, at the lighthouse of Gollenberg. In 1669, Braun used a cast steel reflector with an oil lamp at the lighthouse of Landsort, Sweden (USLS). American Civil War searchlights used crude mirrors made of an unspecified metal that absorbed one-third to one-half of the incident light (Nerz 713).

  Reflector shape. The ideal shape (figure) for a reflector is parabolic; if the light source is at the focal point, then all of the reflected rays will be parallel to the optical axis of the reflector. There were occasional experiments with spherical reflectors at lighthouses, since the spherical shape was easier to achieve. These proved to provide little concentration (USLHS).

  For the techniques of grinding a mirror to a parabolic shape, see Cooper, "Seeing the Heavens" (Grantville Gazette 14),

  Reflective Material. The ideal reflective material would be highly reflective across the visible light spectrum, easily formed into the parabolic shape, resistant to corrosion (tarnishing), easily cleaned and polished, low in density, and inexpensive. Most modern mirrors are composites—typically a metal coating on a glass or plastic substrate.

  For metals, the reflectivities at 400 (blue) and 700 nm (red) are as follows: gold* (39%, 96%), copper* (51%, 95%), silver* (87%, 97%), aluminum (92%, 91%), iron* (48%, 54%), tungsten (46%, 52%), tin* (75%, 83%), chromium (69%, 64%), and rhodium (76%, 81%). Only the asterisked metals are known to European metalworkers at the eve of the RoF. Plainly, silver and aluminum are the best from a purely optical standpoint.

  Silver of course is expensive and so there is some advantage to combining the high reflectivity of a silver coating with a lower-cost metal. A silvered copper parabolic reflector was fitted to the La Heve lighthouse in 1781 (Marriott 25). Robert Stevenson combined an Argand lamp with a silver-clad copper parabolic reflector and, installed at the Bell Rock lighthouse in 1811, it produced 2500 candlepower (USLHS).

  Silver, however, is subject to tarnishing as a result of hydrogen sulfide in the atmosphere (or in perspiration if the mirror surface is touched). The resulting silver sulfide is black. The tarnishing is more rapid if the air is humid.

  Costs could be reduced further by use of speculum metal (45% tin, 55% copper). Its reflectivities are 63% at 0.45 and 75% at 0.65 (Tolansky). Unfortunately, it, too, tarnishes, and it is also somewhat brittle.

  The first telescope with a parabolic mirror was built by Hadley in 1721. It was a six-inch diameter piece of speculum metal. The Royal Society praised his achievement, but expressed the hope that someone would either figure out how to keep the metal from tarnishing or how to make a silvered glass mirror (Pendergrast 161). This proved to be a difficult proposition, and speculum continued to be used well into the nineteenth century.

  When a metal mirror needed to be cleaned it also had to be repolished and often refigured. The Rosse telescope (1845), the largest in the world until 1917, had two six-foot speculum mirrors, o
ne would be in use while the other was being refigured (Pendergrast 176-80).

  ****

  For those for whom cost was an issue, Fitzmaurice invented platinum-glazed porcelain reflectors. They cost one-quarter of the equivalent silvered metal reflector but were inferior in performance. They were used at Sunderland Lighthouse (1860).

  ****

  Various methods of "silvering" glass were discussed in Cooper, "In Vitro Veritas: Glassmaking After The Ring Of Fire" (Grantville Gazette 5).

  Down-time glass mirrors weren't actually silvered; rather a tin-mercury amalgam was applied to the rear surface of the glass. After 1732, James Short tried and failed to use this method to make a paraboloid mirror; he switched to speculum metal (Pendergrast 161). In 1788, Rogers made lighthouse reflectors of "silvered" glass, but they proved to be too fragile USLHS).

  Advances in the arts of silvering glass and of grinding glass to paraboloidal shape made possible the silvered glass paraboloidal mirror.

  In 1835, von Liebig discovered how to deposit pure silver on glass by chemically reducing (with sugar) a boiling silver nitrate solution. Drayton patented several cold processes in the 1840s but the mirrors so manufactured were unsatisfactory (e.g., developed brownish red spots after a few weeks—"measles!") (Chattaway).

  Liebig came to the rescue in 1856 with the first truly satisfactory method, which used caustic soda and ammonia to accelerate the reduction. In 1856, Steinheil used it to silver a four-inch diameter telescope mirror (King 262). Foucault likewise made a silvered glass receptor in 1857, but used one of Drayton's silvering methods (Chattaway). There were further advances in the silvering art that came later (Common). One such was Cimeg's (1861), with Rochelle salt as the reducing agent. EB11/Mirrors describes the Brashear method (1884) in great detail.

  In 1858, Foucault devised the knife-edge test, which could be used to determine how much a glass surface departed from spherical. Hence, you could make an accurate paraboloid surface by an iterative hand grind-and-check process. The same year, he made a 40-centimeter silvered glass paraboloid telescope mirror. The method was perfected by Draper in the 1870s, who preferred the Cimeg silvering process (Lemaitre 20).

  Nonetheless, governments contented themselves in the 1880s with inferior catadioptric reflectors of the Mangin type (see below) for military searchlights (Burstyn). In 1885, Schuckert "invented a machine that could accurately grind glass into a parabolic" curve (USLHS) and quickly put this to work in making searchlight mirrors. These Schuckert searchlights were used in 1887-8 in the Italian campaign in Ethiopia (Rey 97), and a Schuckert searchlight was exhibited at the 1893 Chicago World's Fair. Schuckert mirrors of 30-inch diameter were used to make forty million-candlepower searchlights for the Heligoland lighthouse in 1902.

  Articles in the electrical and military literature credit him with being the first to make "paraboloid glass mirrors with a sufficient degree of accuracy for searchlight work" (Murdock 359). Were they simply ignorant of the existence of telescope mirrors of that type? Or was the hand-grinding done by telescope makers prohibitively expensive for military and lighthouse use?

  In 1909, the mirror alone for Lowell's 42-inch reflector cost $10,800 ($209,200 is the 2001 equivalent) (Cameron 117); a Model-T Ford in 1910 cost $950 (135). (It is conceivable that the high price was necessitated by the degree of accuracy demanded for astronomical work, rather than the hand-grinding.)

  What about tarnishing? On a telescope, the silvering must be applied to the front surface, to avoid ghost reflections from the glass. Hence, the silver is exposed to the atmosphere. It does tarnish, but it was discovered that the old coating could be removed and a new one applied without loss of the parabolic figure.

  On a searchlight reflector, the silvering can be applied to the rear surface, where it is more protected from the atmosphere. However it will still deteriorate with time.

  With large carbon arc searchlights, the heat generated may be such that one cannot use ordinary glass, but rather thermal shock-resistant borosilicate glass (Pyrex).

  ****

  In NTL 1636, aluminum may be available, but only in experimental quantities. For the necessary raw materials and processes, see Cooper, “Aluminum: Will O' the Wisp" (Grantville Gazette 8).

  Aluminum is highly reflective and only a little denser than glass. Aluminum reacts with oxygen in the atmosphere, but the resulting aluminum oxide is clear and hard, protecting the aluminum from further attack. A mirror was first aluminized in 1932 and an aluminized glass reflector was first used in a telescope in 1935. Aluminization of glass requires a high vacuum, but the film is more durable (Yoder 62). Mirrors may also be made entirely of cast aluminum (264).

  For the sake of completeness, I note that other metals have also been used as reflective coatings. Rhodium plating has been used for dental mirrors and chromium for the rear view mirrors in cars.

  ****

  A continuing concern with silvered (or aluminized) glass searchlight mirrors was vulnerability to breakage—the enemy had a tendency to shoot at searchlights. Two types of coated metal mirrors were tested in World War I; one had its coating destroyed after a few hours exposure to the carbon arc, and the other was of inferior illuminating power to a silvered glass mirror (Baird 10-11).

  In World War II, we had 60-inch, 800,000 candela carbon arc searchlights that used a rhodium-plated parabolic mirror (Wikipedia/Searchlight).

  ****

  Segmented reflectors. Hutchinson built faceted reflectors in 1763-77. Some of his designs were tin plates soldered together, but the largest, twelve feet in diameter, was of wood with pieces of mirror glass (clear glass coated with a tin-silver amalgam) attached to approximate the parabolic shape. It was coupled to an oil lamp and reportedly could be seen ten miles away.

  Another glass-faceted reflector was produced by Walker (18-inch parabolic reflector for the Old Hunstanton Lighthouse, 1776). The facets were set in a parabolic plaster shell in a metal frame. Reportedly, its beam of 1000 candlepower was two-thirds the intensity of a one-piece parabolic reflector of the same diameter. Thomas Smith similarly built an 18-inch parabolic reflector with 350 pieces of mirror glass. Used with a lamp having four rope wicks, the combination produced 1000 candlepower at the Kinnaird Head lighthouse in 1787.

  A modern twist on this old idea would be to use spin-casting to create the shell. In essence, when a liquid is spun, its surface takes on a concave paraboloid shape because of the combination of the gravitational and centripetal forces acting upon it. All we need, then, is a substance that will harden into that shape. Appleyard reports that both gelatin and melted wax work. De Paula used plaster. The resulting figures are adequate for solar heating, and hence also for searchlights.

  Spin-casting can be used in place of grinding to create glass paraboloid mirror blanks for telescopes, but you need a rotating furnace, and the molten glass must be cooled slowly (over several months). For telescope use, there is further milling and polishing to make the surface as accurate as possible (Mirror Lab). We don't need this!

  ****

  Lens. Big telescopes use mirrors rather than lenses of the same diameter because the latter are much more expensive. However, Fresnel invented a lens composed of separate concentric annular sections, whose surfaces approximate that of a simple lens of the same focal length. Since it is only using the part of the glass that contributes to the proper refraction of the light, it is much lighter and less costly than a simple lens.

  The more sections there are, the less degradation in performance relative to a one-piece lens, but the greater the cost. The sections may have curved (better concentration) or flat (cheaper) surfaces. A Fresnel lens was first used in a light house in 1823 (Wikipedia/Fresnel Lens). The largest ("hyper-radial") had a height of 148 inches and weighed 18,485 pounds. For a ship's searchlight we would probably use one of "third order" (62 inch height, 1984 pounds) or smaller (USLHS).

  ****

  Mirror-Lens Combinations. Robert Stevenson invented (1849) the holophotal reflec
tor. This combined a central spherical reflector, a peripheral parabolic reflector and a Fresnel lens, and the point was to capture essentially all of the light from the source (USLHS).

  Mangin reflectors were invented in 1876 for use with the carbon arc (Navy 1). This was a lens having two concave surfaces of different radii, the front surface having the shorter radius, and the back surface having a reflective coating (thus constituting a spherical mirror). The radii were chosen so the spherical aberration produced by the lens was exactly opposite to that produced by the reflective coating.

  The Mangin reflector had the disadvantage that it had a longer focal length and therefore a smaller effective angle than a parabolic mirror of the same diameter; if the diameter were 60 centimeters, the angles would be 83o and 123o respectively, and as a result the parabolic reflector would gather 2.11 times as much light (Nerz 715) .

  ****

  Weight. Can a NTL 1630s ship accommodate the weight of a searchlight and its power source? Insofar as the steam engine (including boiler) is concerned, I discussed the issue a bit in "Airship Propulsion: Part Three: Steaming Along" (Grantville Gazette 43). The big uncertainty is the weight of the condenser. For use on shipboard, bringing down the weight of the condenser is less critical, so let's just say six pounds per horsepower—that's 102 pounds for the 17 hp steam engine postulated above.

  I don't have figures for the weight of a 24-inch searchlight, but for a sixty-inch one (delivering 800 million candlepower!), with the six-cylinder gasoline engine, 16.7 kW generator, carbon arc, metal mirror, protective glass, and aiming apparatus all mounted on a small four-wheeled trailer, the combined weight was six thousand pounds. (Fort Macarthur). That may seem like a lot, but it was not unusual for a mid-nineteenth century naval gun to weigh 150-200 times the weight of its shot (Ward 30), which would make the 60-inch searchlight equivalent to a 30- to 40-pounder. (And in the late seventeenth century guns were heavier, 175-250 times shot weight (Glete 516).) If weight scales with beam area, then the 24-inch would weigh only 1,000 pounds, and a 12-incher would weigh 250 pounds.