As mentioned previously, the word "GYRALITE" refers to the gyrating warning lights formerly made by The Pyle-National Company. The illustration of Gyralite Types shows the styles and numbers of some of these.
The 1958 Pyle-National catalog mentions that the intensity of the beam is not as important as the beam pattern. The sealed beam lamp they cite has 200,000 candlepower. Two of these lamps will give a sum of 400,000.
A 32 volt 250 watt headlight lamp with a 16 inch Alzak reflector has 385,000 candlepower. This same headlight with a 16 inch silvered glass reflector produces 752,000 beam candlepower. see below (20750)
"Alzak" is a licensed trademark of
ALCOA (Aluminum Company Of America). An Alzak reflector is made of
anodized 
aluminum. Some advantages
are its light weight, dissipation of lamp heat, less fragility, and brilliance
of reflected light afforded by the Alzak anodizing process. The main
disadvantage to using a silver glass reflector (or a plated metal reflector)
was that resilvering (replating) was required every 2 to 3 years when used in
close range to high wattage bulbs. The process was simple and not that
expensive, but perhaps inconvenient to the railroad industry. The anodized
aluminum reflectors (Alzak) are more durable than glass at close range to high
wattage lamps. The silvered glass reflector, however, will produce increased
reflectivity, and hence candlepower.
The 1950-1952 Locomotive Cyclopedia Sec. 17 Page 653 - “Alzak Reflectors for Headlights” states: Alzak specular-finished reflectors for locomotive headlights have a highly reflective surface that does not tarnish and requires practically no cleaning to protect the brightness. When cleaning does become necessary, soap and water restore the original brillance.
The table above was taken from the Mechanical Engineers' Handbook edited by Lionel S. Marks, 5th edition, 1951. The date is appropriate for the materials used in this era. It is interesting to note the reflectivity percentage of Alzak relative to the materials listed. These values of reflectivity are for flat surfaces. Using these materials in the form of reflectors could result in the beam candlepower outputs like those expressed above for the 16 inch reflector.
US
Pat. 2,607,838 illustrates a reflector Gyralite which used a hinged
hemispherical colored glass unit that could be put into the path of the beam.
Trans-Lite said that one person remembered a version of this light that was
produced and marketed. 
Pyle-National's 20350 Gyralite normally gyrated white light counterclockwise and red light clockwise (as one faced the oncoming locomotive). One can verify this by looking at the illustration in US Pat. 2,607,839 of the screw type shaft which advanced a red colored globe over the lamp. This light was marketed as a built-in light for diesel locomotives. It was used on the Southern Pacific - Fairbanks Morse "Train Master" locomotives as well as others. The sweep of the beam was a standard 55 to 60 per minute.
Manual control settings for these lights included: ON, OFF, DIM, FULL, INCH, and RED. The "Inch" setting enabled the engineer to ease the light into a desired setting.
It appeared to be a dilemma when these lights first were put into work on how to use them as regular headlights without the gyrating motor on so that the light beam would be projected out parallel to the tracks automatically, without having to be "inched" back into position. A braking device on the gyrating motor's shaft was utilized (illustration : US Pat. 2,666,193) to return the light to its normal position on terminating its gyrating function. Gyralites utilizing a reflector are not currently being produced or marketed by Trans-Lite, Inc.
Some of the other rotating
reflector versions were the 10350 Gyralite and the 10590 Gyralite. Both of
these used 10 inch reflectors and P-25, 250 watt, 32 volt bulbs.
The 10350 type could be obtained in two forms. One form was adapted for permanent mounting to a locomotive or railcar while the other type was adapted for portable use. This second form had a hanger with a clamp arrangement on the rear of device which could be mounted on the vestibule gate of a rear car of the train or, as used by Canadian Pacific's RDC units (Rail Diesel Cars produced by Budd), on the forward ends of these motorized railcars.
One of the very few oscillating warning light applications in Canada was this portable single white light Gyralite used by Canadian Pacific on their Budd RDCs until the VIA Rail Canada takeover in 1978. Wheeling & Lake Erie mounted a couple of these Gyralites on their Berkshires before settling on the clear lens, sealed beam version of the R-250 Mars Light. (see below).
Kevin Holland
The 10590 was adapted for flush mounting to a locomotive or railcar.
A low voltage (6 or 12), light weight, compact unit, the 15100 Gyralite was marketed using the same type of rotating reflector mechanism.
Two types of sealed beam bulbs used in Gyralites are the PAR-46 and the PAR-56. The notation "PAR" denotes a parabolic configuration of the sealed beam's reflector. The number following this notation refers to the diameter of the bulb in 1/8" multiples. A PAR- 46 bulb therefore has a diameter of (46x1/8") 5.75 inches. A PAR-56 bulb has a diameter of (56x1/8") 7 inches. These bulbs are classified as "narrow beam". The lens doesn't have the longitudinal rectangular pattern which aids in spreading the beam as in an automobile headlight.
Specifications for GE bulbs:
PAR-46: 32 Volts/150 Watts/1950 Lumens/100,000 Beam-Candlepower
PAR-56: 30 Volts/200 Watts/3700 Lumens/270,000 Beam-Candlepower
There are Gyralites models that are essentially identical except for the
particular bulb diameter they accomodate. The 17340 Gyralite uses the PAR-46
bulb whereas the 17350 Gyralite uses the PAR- 56. This type of Gyralite uses 2
sealed beam bulbs. The 17340 Gyralite has 2 plates to which each bulb is
mounted. One plate is stationary and the other is oscillatable. The 17350 Gyralite
has the stationary sealed beam bulb mounted in the unit's door. The other bulb
is mounted on an oscillatable plate. Both these light units were marketed as
"combination back-up types", apparently for use at the rear of
railcars. (There was also a 17385 Gyralite marketed for this application.) 
Trans-Lite produces the FG-5031-1 light which incorporates the physical and mechanical aspects of the 17340/17350/17385 lights. This unit is used on Bessemer & Lake Erie locomotives where it is mounted vertically between the number boards with the top bulb being capable of oscillation. The FG-5031-1, which looks like the non-flanged 20585 Gyralite, is also installed with the hinges towards the left side of these locomotives.
The 17540 Gyralite uses the PAR-46 bulb. The 17550 and 17570 Gyralite both use the PAR-56 bulb. Other than the location of the conduit holes on the back of each, they are identical. The 17550 Gyralite's 1-1/8" conduit entrance hole is located in the upper half of the unit on the vertical median (as one faces the back of the unit in its running-mode position). The 17570 Gyralite's 1-1/8" conduit entrance hole is located in the lower right quadrant of the unit near the perimeter (as one faces the back of the unit in its running-mode position).
The 17540 (or similar) (see illustration photo) single light gyrating units can probably still be found on some SP locomotives. These are permanent or portable units and are available to run on a variety of voltages. On one unit tested, the beam rotation was 93 rpm.
The 175x0 series were usually back or surface mounted. The Frisco GEs (BN on merger) used a recessed unit mounted in the locomotive's nose. These recessed units had a flange. The Frisco (BN on merger) ran with a clear "lens".
The FG-4509 Gyralite is like that of the 17550 or 17570 unit but this unit does not have a case. It was built for interior mounting. This unit is mounted behind a glass shield in the Metra cab cars. Metra oscillates a clear beam with this unit.
The 20585 unit described below had rotation rates
of 50-65. Restriction of the light mechanism by dirt buildup, tight belt
tension or compression on the joint that connects the slide assembly to the
bulb plate (via nut) will effect loading as well as rpm of the beam. These
factors may also cause an increased current draw by the motor with accompanying
overheating. 
The popular 20585 Gyralite (see illustration photos) uses twin 30 volt PAR-56 sealed beam lamps, each having a rating of 200,000 candlepower (GE specs.: 270,000). This light unit can be run with both lamps illuminated white as Caltrain and MARC (Maryland Rail Commuter Service) does or red using red lenses. They can also be run as one white and one red. In this second instance, only one lamp would be on . In the case of a vertically mounted 20585, the lower lamp lens is the one usually chosen to be red.
I made an inquiry to Kopp Glass, Inc. on the type of glass used in the lenses of the 20585 Gyralite. (It appears from the etched markings on the inside rims that they made the glass lenses for the Gyralites as well as the Mars oscillating lights.) This was their reply:
The circular roundel which you mention (lens in 20585 Gyralite) is not made of Pyrex. Pyrex is a patented product of Corning. We use borosilicate glass in our products. Our glass does meet AAR specifications for heat resistance but it does not quite meet the heat tolerance of Pyrex.
Darin M.
Bernardi - sales engineer
Kopp Glass, Inc.
2108 Palmer St.; Swissvale, PA 15218
It does appear that the 7-9/16" roundels were also made by Corning as evidenced by the "PYREX" trademark on the inside rims of ones I ran across.
Pyle-National established the precedence of the positioning of the "red" lens when used in a clear-red lens combination. Their literature (together with Trans-Lite's) on the 20585 consistently illustrates the "red" as being the lower of a red-clear (red-white) combination.
Pyle-National - Bulletin 1254-G
dated May 10, 1958
Gyrating Visual Warning Signals
Type 20585 -Sealed Beam - for Vertical Mounts in Diesel Locomotives:
"Red-white combination types have the red roundel in the lower
position."
The 20585 was furnished (or could be ordered) with red lenses giving one the option, however. Specifications on the 20795 Gyralite show the "red" light of a red-clear (red-white) combination as being the lower of the two. No option was given in this case. Specifications for the 20775 (see below) shows a 3 light combination in a triangular pattern. Here the top bulb is the one which is "red". Again, there was no option for alternatives.
Originally, a 32 volt motor was utilized in the 20585. A 60 volt version was developed along with the present 74 volt for diesel compatibility. The light unit is provided with 74 volts and through a dropping resistor, the lamps and motor (if necessary) both receive their correct voltages. There are, however, railroads that apparently wired the two 30 volt bulbs in series without a dropping resistor. An adjustable motor armature voltage dropping resistor was used to obtain the desired 55-60 revolutions per minute of the beam. (This was cited in a 1950s catalog, but is apparently no longer used.)
At present, the railroads using gyrating warning signals are mainly confined to: Kansis City Southern, Caltrain (presently removing - 1999), MARC, and Chicago's Metra. They all appear to use the same 20585 Gyralite which is now sold by Trans-Lite, Inc. This same type gyrating light was used by the Southern Pacific before the overall decline of gyrating warning light use.
The EMD part number of this unit reflects its many
options. Caltrain used a Gyralite with an EMD #8279365. This number is
equivalent to Trans-Lite's number FG-20585-DWFV.
F = original paper size (17"x22") or 4 - 8.5"x11" sheets
G = Gyralite product line
20585 = basic drawing number
D = Dual light
W = White
F = Flange (around body of unit for mounting)
V = Visor
The 20585 was available in 2 mechanisms. photo One was for vertical mounting (one bulb on top of the other) while the other was for horizontal mounting (bulbs side by side). The 2 mechanisms were produced to assure that the lowest portion of the beam sweep would be equivalent to a stationary headlight, that is a light beam which is parallel to the tracks. The Beam Patterns page shows the respective patterns for each mechanism. The position of the slide bearing as well as the slide arm length, together with the alignment of the pulley to which the slide arm is attached, determines the beam pattern generated. In the Vertical 20585, the pulley to which the slide arm is attached is on a line connecting the center of each bulb. In the Horizontal 20585, this pulley is displaced off this centerline.
It appears that because of maintenance costs or whim that crossing lights (ditch lights) have become prevalent on most lines. Unfortunately, they are not as spectacular in operation as the gyrating warning lights and are therefore not nearly as effective.
Southern Pacific (as well as IC/ICG and FEC) used what is known as an "Oscitrol Headlight" as a replacement for some their mechanical gyrating warning lights. This warning light is explained in the Oscitrols page.
As far as the contemporary mounting location of the 20585 gyrating light units is concerned, I notice that it is (or was) preferential to each railroad. On the Southern Pacific, Caltrain, MARC, and Chicago Metra units I have seen, the roof seems to be the favorite for the twin lamp units. On the former Denver & Rio Grande Western and the present Kansas City Southern, the nose seemed to be the preference. (see illustration photos)
Some railroads mounted their gyrating lights of the form in illustrations US Pat. 2,677,121 to the plate between the number boards. The unit's back end was just butted up against this plate and secured to it. (see illustration photo) In viewing pictures of these lights on these locomotives it can be noted that these contemporary gyrating lights are of the twin sealed lamp units mounted on a plate which is rotated in an elliptical pattern. When gyrating, both the headlamps are undergoing the gyration. On a locomotive one could have both lights on and reinforced to form what appears to be a single powerful beam, if desired, or one of the lamps could have a red screen or lens. This way one of the lamps could be run white and the red one used for emergencies. An illustration of this type of light could be best shown by the illustrations of US Pat. 2,677,121.
The compact size of the 20585 Gyralite makes it the most practical twin beam unit. A V-belt drive in the unit's drive train make for a more compact motor location. The V-belt drive also acts as a shock absorber in the drive train thereby contributing to the smoothness of the beam's rotation. Apparently the gearbox design was used with both the original as well as the modern motors. The gearbox reduction was 80:1.
The relative position of the slide bearing bolt to the center of the pulley on which the slide assembly is attached, as well as the geometry of the mechanism, causes what is known as the "delayed sky effect" in a vertically mounted unit. This causes the beam to traverse at a slower rate on the top (skyward) portion of the elliptical sweep than the lower (faster) portion. This is described in detail on Beam Patterns page. The depth of this unit is 12-1/8 inches (nearest 1/8 inch) .
The dual light unit illustrated in US Pat. 2,595,253 apparently had a circular beam pattern. If this light was produced, the market was short lived, as no photos seem to show this model. Trans-Lite states that they do not have any information on this unit. On the takeover of the lighting line from Pyle-National, their main interest was in the lighting units that were in production. If this unit was indeed produced, they were not aware of it.)The vertical 20585 seems to have been preferred for its large elliptical beam sweep, together with other design advantages.
Also it should be noted that a patent (US Pat 2,607,838) for a
reflector Gyralite was issued in this era that used a hinged hemisphere of
colored glass that could be positioned over the bulb to produce a colored or clear
beam of light. A person at Trans-Lite remembers a version of this unit being
produced. 
The illustrations of US Pat. 2,654,875 illustrates a similar arrangement in a round casing with a color screen over one of the sealed beam lamps. This design was incorporated in many of the streamlined diesel units of the 50s. Some are mounted with the two lights vertical while others have the two lights horizontal. This patent is listed as a basis for many of the plate mounted sealed beam Gyralites.
It should be noted that the 175x0 series (above) also came under this patent. This design was the predecessor to the US Pat. 2,677,121 on which the 20585 Gyralite was based.
A more recent patent, US Pat. 2,780,796 illustrates a combination of a reflector and moving plate type warning light. There were two models manufactured under this design. The 20750 Gyralite used a reflector made of silvered glass while the 20770 Gyralite used one of Alzak.
This patent described a method of mounting a sealed beam lamp(s) to the back of a plate. This plate could be the support for a reflector as in the 20750 or 20770 Gyralites, or could be the support for an array of sealed beam lamps. The latter case is represented by the 20795 Gyralite in which 2 bulbs are secured to the plate, the bottom utilizing a red lens, or the 20775 Gyralite in which 3 bulbs are mounted in a triangular fashion with the red bulb at the triangle's top or apex. (IC and C&O were users of the 20775.)
All the Gyralites of this type were promoted as having very compressed beam patterns for maximum coverage in the horizontal direction. Pyle-National's Bulletin No. 1254-G (1958) states that in such a compressed elliptical pattern, the observer would actually see two flashes of the rotating elliptical beam per cycle, giving the impression of a rotational speed of 2 times what it actually was. On observing the operation of the 20750 Gyralite on the SP E9A at the California State Railroad Museum, the beam pattern was compressed to the point of looking lateral.
The 20750 or 20770 Gyralites both used a PAR-56 sealed beam headlamp mounted to the reflector (may incorporate a "red" screen or lens) which was capable of oscillation. In addition, a P-25 bulb usually "clear" (or white) was used to act in conjunction with the reflector. This device may be operated in the oscillating mode by selectively energizing either of the light sources. The unit can also serve as a headlight in the stationary mode using the "clear" (or white) bulb selection, if desired. The 20750 Gyralite was promoted for the most power at the best efficiency:
Beam candle power of white light:
In using a 16 inch reflector with a 32 volt 250 watt bulb, this unit would
produce 750,000 beam candlepower. [This was considered very efficient as the
most 2 sealed beams in a 20585 could produce would be 540,000 (2 x 270,000).]
Bulletin 1254-G dated December 1, 1958
Sheet No. 17
The 18350 Gyralite was not used
on locomotives. Rather, it was designed for flush mounting on railway cars. US Pat. 2,645,761 illustrates
this unit. It used a P- 25 bulb and a 10 inch reflector for the gyrating
portion. A PAR-56 sealed beam headlamp was used for the stationary part. This
stationary light served as a backup light for the train, not only warning the
crew that the train was backing up, but also illuminating the track for the
trainmen. A color lens (usually "red") was used in the gyrating part
of the device. 
The 15360 Gyralite is a beacon type unit. US Pat. 2,846,663 illustrates this unit. This beacon is included in this page due to its "Gyralite" nomenclature. The beam from a PAR-46 sealed beam mounted in the units top is reflected at right angles using a dual inclined plane reflector. This reflector revolves at 48 rpm, producing 96 flashes per minute. Coverage is 360° in the horizontal plane of the beams rotation.
This Gyralite was marketed for industrial and locomotive use. It is also discussed on the Beacons page and Strobes page.
"Ash-Can" Gyralites:
March 1998 "Trains" Magazine:
Cadillac style
Electromotive SD7's and SD9's were among SP's most beloved locomotives
by Ted Benson
Page 60 has 2 photos of SP 2742 SD7 "Black Widow" with a large cylindrical light structure with base of unit mounted at the intersection of the locomotive's front and roof. The bottom of this cylindrical unit was on a plane of the forward projection of the roof. (unit mounted above the roof by a distance equal to the diameter of structure.) These were called "ash can" signal lights due to their large diameter and cylindrical appearance. Gyralites such as the 20350 and 20750 as well as others were available for use in such mountings.
Page 61:
But time was creeping up on the pioneer SD's. Upon casual inspection, most of
the changes seemed cosmetic. Black Widow paint gradually turned scarlet-and-gray,
"ash can" signal lamps gave way to smaller Mars and Pyle-National
units, and a 1965 renumbering moved the units from random placement in the
5200-5300-5400 series to an organized grouping in the 2700/- 3800-3900
sequence.
Gyralite - rotation direction - viewed from cab of locomotive:
The catalogs showing the 32 volt motors used in the 20350 and 20585 show 4 wires supplied to the motor. Selectively connecting a field-brush pair to one polarity would produce the desired direction.
The 20350 Gyralite (see below) had a threaded shaft which was cut so that counterclockwise rotation of this shaft would produce a clear or white beam. The majority of Gyralites I have observed (175x0/20585) seem to have the clockwise rotation. The Pyle-National Bulletin 1254-G page 2 shows this clockwise direction in its illustration with an arrow.
Pyle apparently went to a 2 wire supply system in their later motors. In those of original design, which had a field winding, the field was connected with the brushes in series. Such a wiring scheme made only one direction of rotation possible, regardless of supply polarity. This trend was also seen in the motors used in the Mars Lights.
A SB-WR-2-200 32 volt Mars Light had a cooling fan on one end which dictated one direction (pitch of the fan blades). It appears that the direction of rotation in the 2 wire field winding motors was intended to be this clockwise direction.
The latest generation of motors has the field winding replaced with permanent magnets. Using a 2 wire scheme with this motor allows the directional rotation to be easily selected by the polarity of the voltage connected to these wires. The Kansas City Southern (KCS) Railroad illustrates this, in that locomotives can be seen having both clockwise and counterclockwise rotations (20585). This was an option in installing the wiring to the motor in most units. Installation instructions gave the installer the option.
Another aspect of the 32 volt systems was that both the 20350 and the 20585 Gyralites appeared to have used identical motors and gearboxes (shown as identical in catalog drawings). The gearboxes were double reduction, giving an overall reduction ratio of 80:1. This yielded the 55-60 rpms of the beam that is quoted in the catalog literature.
Horizontal Mounting
A mounted 1750x series has the hinges at the bottom. This is also the case in the 17340, 17350, and 17385 Gyralites. The 20585 mounted horizontally has the hinges at the top. One possible explanation for the hinges at the bottom may have been for maintenance. The door would need to be secured to replace the bulbs or perform other maintenance on these units. The disadvantage of having the hinge at the bottom is that if the cover isn't secured with care, the door could drop open during travel and the crew would not be aware of this. The colored lenses were mounted in the door, so this could hinder the warning of another train if an emergency occurred. The 20585 mechanism could slide out of the mounted case for maintenance, allowing the hinges to be at the top. As stated above, the carry over for a red roundel on the bottom of a vertical 20585 is having the red roundel on the left of a horizontal 20585 (hinges at top).
The 17340, 17350, and 17385 Gyralites are shown with the oscillating lights on the left of the unit, but the hinges are at the bottom. It appears that the doors for these 173xx series Gyralites may have fit the 20585 cases. This could have caused locomotives to be observed as having the red roundel on the right in a horizontal 20585. Trans-Lite verified this, stating that the doors are interchangeable so it is the lens installation that dictates which door it is (in a red-clear unit). Deviations from specification data on a Gyralite would therefore result if the lens was moved or if the doors were used interchangeably without taking the stated location of the red lens into account.
Headlight Louvers:
Headlights mounted between the numberboards use a visor to keep the glare off of the top of the nose. This is also why railroads paint the tops of the nose black or some other dull color. These visors are made of sheet metal and after years of use (read abuse) such as cleaning of the lens, or rust, they tend to bend and break off.
This is also the same for locomotive headlights. The roof mounted units have visors but over time they may lose them for the same reason as listed above. The headlight may also be replaced with a standard type without the visor. GE uses a thicker sheet metal on their visors and they do not use the intermediate shields in the visor so it makes cleaning the headlight easier, therefore they seem to keep their shields longer. EMD uses a multi-tiered visor similar to the Gyralite visor that makes it harder to clean the bulb/lens so therefore the visors take a lot of more abuse.
Another reason you may see locomotives without the visor is replacement parts. Since railroads do not keep every variation of the headlight/Gyralite around at every remote servicing location, they use what they have on hand to replace broken parts. There objective is to keep the trains moving. Therefore you see replacements without the visors applied.
Bill Kaufman
KCS is an unusual railroad in many respects, the Gyralite being only one (albeit very visible) example of their willingness to be "different" from the pack. I have never had to work on an ex-KCS Gyralite, only because their policies on locomotive selling/swapping/scrapping keep them from being on the market like units from other roads... unlike ex-SP examples.
Because most lights I get to play with have been in service for a long time, they tend to be pretty rough. In my experience, the biggest problems with bad order lights (Mars included), is caused by congealed or dirty grease in the bearings/floating ends. It is no wonder that the motors go bad when they are subjected to loads way over what they were designed to take. I agree with your point about the motors; I think that Trans-Lite has improved them greatly by going to the "can-type" motor... they are sure easier to deal with when you have to pull that angled drive off the front! As far as sealing goes, I haven't seen enough real recent ones to comment on that, but I can state that a lot of lights suffer from water (and other substances) coming through the back drilled conduit holes... SP liked to skimp on the mounting gaskets.
Griffin Hamilton
When the KCS first acquired the Trans-Lite system the two 30 volt bulbs were
wired in a series connection and used a voltage drop resistor for dimming
purposes only.
After approximately two years of operation, the KCS changed to a parallel system because with the series connection, both lights would go out when one bulb became defective. The modification does require the use of 2 voltage drop resistors along with the dimming resistor.
KCS Mechanical Department (1999)
Trans-Lite has assemblies that allow 2 lamps in dual fixtures to be wired in series and have a resistor to bring the locomotive voltage to 60 volts. These operate by kicking in another set of resistors when one lamp fails in order to keep the voltage on the remaining lamp at 30 volts. This is called the "Continuelite Control" and it is used more on commuter cars with a dual headlamp over the end door.
Trans-Lite has made improvements on the 20585 Gyralite over the years as recognized by the newer units utilized by KCS. The motors supplied to Trans-Lite have been redesigned to increase their longevity. The unit has also been made more environmentally sound thereby increasing resistance to the influx of sand and other detrimental substances.
___________________________________________________
Photos on this page were taken from
Pyle-National catalogs
