Facebook Page
Electrical Aspects - Gyrating Warning Lights


Electrical Aspects



The locomotive electrical system is full of anomalies. For one thing, no agreement can be reached on the accessory power supply voltage. Hence we see values of both 72 volts and 74 volts being used for the charging voltage. Secondly, it may appear that the bulb designers were alienated from the needs of the locomotive electrical system but it was the strength of the bulb's filament that was the obstacle.

The steam era used mainly the 32 volt electrical system (other voltages also used - turbogenerators). The diesel system implemented a 72/74 volt system. It appears that in this diesel system everything but the headlight bulbs were run on the 72/74 volts.

The bulbs used in the first diesels were 32 volt as well as 12 volt P-25 screw in bulbs.

The 12 volt systems used a dynamotor (motor coupled to a generator) for a power supply (high current). Pyle-National promoted upgrading these systems using 32 volt P-25 or 30 volt PAR-56 bulbs together with voltage dropping resistors.

The designation for the 30 volt 200 watt PAR-56 bulb is:
200PAR - for GE

A 30 volt sealed beam headlamp was developed apparently to replace the 32 volt bulb-reflector combination. Yes, 30 VOLTS. Even the bulb companies (GE & Osram Sylvania) cannot come up with any answer why they called this a 30 volt bulb instead of a 32 volt bulb. These two companies also state that 30 volts is the maximum recommended voltage for these 30 volts sealed beam bulbs. It may seem questionable, however, that it would have been recommended to the railroads to drop 2 volts with resistors. For example, many steam locomotives which were set up to output 32 volts, did end up using the 30 volt sealed beam units (NKP Berkshires with Clear lens option of the SB-R-250 Mars Lights).

It was speculatated from one source that the 30 volt sealed beam was intended to be used with a ballast resistor or resistor "package" which included the dimming function as well as a series resistance to increase bulb life.

As the diesel replaced the steam locomotives, no 72/74 volt bulb was utilized that would allow railroads to divest themselves of the voltage dropping resistors for fifty years. see GE below - "produced in 60s"

It appears that the mechanical strength of the filaments dictated the use of the 30 volt bulbs. A 75 volt bulb has a decreased diameter. The shock to the filament in yard service from banging into railcars as well as vibration on the track causes filament fatigue and breakage.

An analogy of this predicament was the early automobile 6 volt electrical system. This system was used due to the headlight bulb filaments not being able to withstand the shock or vibration in a higher voltage. The current draw was high. Finally the 12 volt headlight bulb was proven and is now the norm.

In the development of the oscillating light (see History page - Mars), the 32 volt standard headlight bulb was later changed to 12 volts bulb due to the fact that the 12 volt bulb filament had increased rigidity and therefore less prone to breakage. Later, companies went back to the 32 volt bulb - apparently improved. Both Pyle-National and Mars were stuck with having to use the headlight bulbs available and had to deal with the electrical inefficiency of the high wattage voltage dropping resistors.

Recently (90s), a 75 volt 350 watt sealed beam bulb 350PAR56/SP was put into the specs. of both GE and OSRAM SYLVANIA.

The "SP" in this bulb's designation means:
SPot - for GE
SPecial/Multi-Purpose - for OSRAM SYLVANIA

As of this writing, I have had one RR who has stated experiencing less bulb life on this bulb. Most likely, strides will be made to improve the bulb - as was the case of the 32 volt bulb.

It should be noted that OSRAM SYLVANIA states NO specific application for their 350PAR56/SP bulb. They are not committing the bulb to railroad use as of 1999. They claim 2 years of research in their bulb development.

GE claims their 350PAR56/SP was produced in the 60s. The GE application specs. calls for the application of "ditch light" for this bulb.

Railroads found that they could use this as a headlight bulb also. According to OSRAM SYLVANIA's beam spread and candlepower rating on the bulb, it meets the Code of Federal Regulations requirement of candlepower and is as good as the beam spread of 30 volt sealed beam bulb in use for headlights. This eliminates the voltage dropping resistors. I was told that the new locomotives had gone from voltage dropping resistors to DC-DC converters, but in view of the use of the 75 volt bulbs, this is not necessary.


The table below shows the values for the 30 volt 200 watt PAR-56 and the 350PAR56/SP [75 volt-350 watt] sealed beam bulbs:

200W 30V PAR-56 Bulbs


beam angle (50% cbcp)

field angle (10% cbcp)



9 x 9 (1)




4 x 4

10v x 9h

350W 75V PAR-56 Bulbs


beam angle (50% cbcp)

field angle (10% cbcp)



8 x 8 (2)




6 x 6

12 x 12


see other max. cbcp values based on tests (below)
cbcp = center beam candlepower

beam angle and field angle = angle which the light output drops off to 50% and 10% of the cbcp, respectively.
[expressed in (vertical) x (horizontal)]

The term "beam angle", "beam spread" and "beamwidth" are used to denote the angle at 50% cbcp.

(1) GE value given in specs. - other values from tests (see below)
(2) GE gave out this value but on checking again, claimed it was unavailable.

The following isocandela information is taken from the Final Report (July 1995) of the US Department of Transportation and the Federal Railroad Administration entitled: Safety of Highway-Railroad Grade Crossings - Use of Auxiliary External Alerting Devices to Improve Locomotive Conspicuity DOT/FRA/ORD-95/13 DOT-VNTSC-FRA-95-10

There were 2 sets of data submitted to the DOT/FRA for the GE locomotive bulbs. The US Coast Guard and the Quest Corporation (North Royalton, OH) furnished isocandela diagrams to the US Dept. of Transportation (FRA division).

The beam widths of the GE 30V 200W bulb were stated on the Beam Patterns page. This will be repeated here also. One can notice the results of these tests and take note of the values above. In the Appendix C of the report cited above, GE has its name on the isocandela diagrams and as seen below in a reply from Quest, these diagrams were apparently furnished to Quest by GE.

GE 200W 30V PAR-56 Bulbs


beam angle (50% cbcp)



5 x 5



4.5 x 4.5 based on interpolation

GE 350W 75V PAR-56 Bulbs


beam angle (50% cbcp)



15v x 10v



6 x 6 - interpolation (nearest deg.)

The maximum candela values were taken from page 4-19 of the above referenced report. The beam angles were taken from Appendix C of the above referenced report. Interpolated values are based on the observed isocandela contours and are therefore estimates based on these contours. The isocandela contours are not perfect circles on these diagrams. There is an apparent averaging that takes place to arrive at a published value. The theoretical perfect parabolic reflector with the filament at the focal point (filament being considered as a point) is not realized. Any biases in the manufacturing of the bulbs will depart from a theoretical "ideal" diagram.

As stated on the Beam Patterns page, I obtained a max. cbcp of 193,000 for the GE 30V 200W PAR-56 bulb with a voltage (under load) of 27 volts. On inquiring to the Quest Corp. on tests results and spec. data on the GE bulbs, they stated:


Quest only passed on to Tony/US DOT-VNTC graphs that were provided to Quest by the lamp manufacture(s). We have no further specific details to provide.

Based on information conveyed to Quest by lamp manufactures and some of our own limited testing and research, the following generally will apply to the PAR-56 lamps:

1. The light output of the lamp is a non-linear with respect to the input power. Light output is non-integer power of the voltage.

2. A +/- 5% change in the input power voltage will generally result in a 20% change in light output. At the lamp's rated voltage this same +/-5% voltage change will generally either half or double the lamps rated life. The 27V input power you stated you used to obtain light output ratings falls in the "MEDIUM" headlight switch setting range used by a number of railroads. 30V headlights in a MEDIUM setting generally translates to a 22-27V range, although typically closer to 22-24V.

3. Over the lamps rated life, ie. 500 hours average, the light output will decrease ~20%, due to bulb blackening related to filament evaporation. When the filament becomes thinner the resistance increases also reducing light output.

4. For lamps with an "average life rating", this typically only means that statistically 55% of the lamps will meet or exceed this service hour rating.

5. Because of the construction techniques associated with making sealed glass incandescent lamps and their filaments, each lamp will vary somewhat from each other and will adhere to their general performance characteristics.


The resistor values of the control to a locomotive headlight are:

Supply voltage = 74 volts
Bulbs: PAR-56   30 volts/200 watts

Primary resistor = 4.7 ohms
Bright = 1.6 ohms
Medium = 1.3 ohms
Dim = 5.6 ohms

Resistors are cumulative
ie Bright = 4.7 + 1.6
ie Medium = 4.7 + 1.6 + 1.3
ie Dim    = 4.7 + 1.6 + 1.3 + 5.6

I dug this up on one of my schematics. According to the GE electrical schematics for the Southern Pacific B30-7's, the headlight voltages are as follows:

Bright = 30V
Medium = 22V
Dim = 14V

Bill Kaufman


The specs were run through a computer program also:

Assuming a bulb resistance of 3.5 ohms and the following intensity settings as specified in your letter ... I get.

Bright  =  Bulb voltage drop 26.43 V, Power 200 watts
Medium = Bulb voltage drop 23.33 V, Power about 156 watts
Dim = Bulb voltage drop 15.5 V, Power about 68.8 watts

Harold Scadden


Many in the field of electronics are appalled on being informed of the circuitry involved in supplying voltage to the headlights and oscillating lights. As discussed above, however, this was necessary because of the mechanical stability of the filaments in the bulbs used. The idea of high wattage resistors and their associated "wasted heat" is in conflict with the new efficient methods of voltage conversion. The problem seems to have been that the bulb manufacturers could not readily come up with a strong enough filament at a standard locomotive voltage end the inefficiency of the electrical lighting systems. This has improved, however, with the new 75 volt sealed beam lamp. Voltage dropping is especially a problem in the single 12 volt headlight units (12 volts @ 480 amps). This bulb would calculate a voltage drop wattage of 2400 watts (dynamotors were used for 12 volt systems instead of resistors). In order to have any longevity, 20% to 100% reserve should be figured into the circuit.

The use of the dual headlight fixture evolved, not because it is required, but because on the loss of a single headlight, specific caution measures had to be performed at crossings. It was apparently conceived that having 2 lamps wired in parallel could avoid having to deal with these measures. The use of the PAR-56 30 volt 200 watt bulbs in a parallel configuration called for dropping 42 volts @ 7 amps (for 32 volt P25 250 watt --- 8 amps). The wattage calculated would be close to 300 watts. It is again suggested to have a wattage reserve over this for increased resistor life. Many of the dual Gyralites and dual Mars Lights had their bulbs connected in series. This helped in the conservation of the resistor heat dissipation, but there was another problem (see below). Connected this way, the calculated resistance would be for a voltage drop of 12-14 volts @ 7 amps or close to 100 watts.

One can notice in the resistor specs.on the dropping resistors for the dual clear and red-clear Gyralites that a 400 watt resistor is listed for both. In the dual clear instance, in which both bulbs are connected in series, a 400 watt adjustable 3 ohm resistor was used. The calculated wattage would be 100 watts. A 400 watt resistor is used. For a red-clear combination, a 400 watt adjustable 8 ohm resistor was used. The calculated wattage would be 300 watts One problem is that the 400 ohm rating is at 8 ohms but the resistor would be set to possibly 6 ohms so this resistor is running close to its limit.


The use of bulbs in series:

For a start the two bulbs won't be identical - one of them will get more of the voltage and will be stressed more. Secondly,  evaporation of metal increases the lamp resistance at a rate depending on temperature (this is the main life-shortening mechanism apart from vibration) so when you have a series pair, whichever one starts out hotter will steadily get more and more of the voltage and get hotter and hotter until it fails.

Anthony New


Since the bulbs are not a "perfect match", there is a tendency for a deviation in the voltage drops between the 2 bulbs. One of the bulbs will have an increased value of resistance over the other. Even though this resistance is small, the bulb with the larger resistance will tend to burn the hottest and as a result, its filament will decrease in cross section. This decrease in cross section will then lead to an increase in the filament's resistance.....the cycle keeps repeating.... the bulb will fail.

I have seen this phenomenon in the dual series wired bulbs in the 20585. One of the 2 bulbs usually exhibits the blackening associated with the vaporizing of the filament, whereas the other bulb will look new.


Resistance rises, so the total current for both bulbs decreases. It is the voltage drops between the 2 resistors, however, that are affected and lead to a bulb failure. The current in a series circuit depends on Ohm's Law, that is, on the sum of 2 resistances.
The power is divided between the 2 bulbs.

Let us take an example. 2 bulbs with 24V 1A nominal each, connected to 48V. [both bulbs are rated at 24 watts]

Assume bulb 1 rises its R to 30 ohm.
Total current = 48/(30+24) = 0.8888A
voltage across bulb 1 = 30 x 0.8888 = 26.6667V
voltage across bulb 2 = 24 x 0.8888 = 21.3333V

power on bulb 1 = 23.7W
power on bulb 2 = 19W

You see, the power in bulb 1 remains rather constant at 24W, but the wire diameter has decreased, so it is no longer a 24W bulb. The decrease in diameter decreases the wattage rating of the bulb even though the current in the series circuit has been decreased (due to the bulbs increased resistance).
It has "become" a bulb which can only tolerate the current of perhaps 19.2W but the power on the bulb is 23.7W. It is overstressed.

Lifetime calculation is done with constant voltage assumption. It is well known that an aged bulb will dim, but this is commonly accepted when talking about its lifetime. In the series connection the "environment" is non typical, unusual ...

With 2 bulbs the effect is not so obvious, but try to calculate it with 3 or more bulbs. Then you will see that one bulb, that changes its resistance - will burn through very quickly.

Another point: when the burnt through bulb is replaced by a new one, the other one will be the "poor" bulb.

If in parallel, then each bulb will reduce its power when its wire gets thinner and so it will extend its life, to the nominal value. But in series connection the bulbs will "die" earlier. That's all.

The additional effect of the temperature of the filament should also be considered. When it is hotter, it has even more resistance, getting more partial voltage and power.

Franz Glaser, Vorderweissenbach Austria


Bulbs in series in dual headlight fixtures:

As for wiring the headlamps in series, I do not suspect this was done or if it was, not for a long time. Remember, the laws concerning the use of a headlight changed towards the end of the steam era. FRA rules mandated that a headlight be illuminated on the train day or night. Therefore, if you have two bulbs wired in series and one fails, according to NORAC rules, the train has to do all of the following:

1. Illuminate all exterior lights. Ground lights, step lights, etc.
2. Notify the dispatcher.
3. Ring bell continuously.
4. Sound horn frequently.
5. Approach all crossing prepared to stop. They can proceed across only at 15MPH.

This is an operational nightmare when you are trying to get trains over the road. This helps explain why the headlight systems are actually two separate circuits from the circuit breaker. Two sets of resistors and two independent bulbs. This way the chances are greater that when one fails they still have the other until they can get somewhere where the system can be fixed or the locomotive swapped.

Bill Kaufman


Use of Resistors in locomotives:

The steam locomotives used a 32 volt power supply to the locomotive accessories. There was therefore no problem (if the 30 volt sealed beam = 32 volt bulb) in dropping voltage to the 30 volt bulbs or 32 volt motors.

The power supply of the diesel locomotive is listed as 72-74 volts. The use of the 30 volt bulbs in the Gyralites requires a resistor(s) to drop the voltage to level which will assure that the bulbs will be supplied with their compatible design voltage. The operation of a 20585 Gyralite (or other dual type) which has a red-clear bulb combination, will require resistors if the bulbs are rated as 30 volts. With the advent of the ditch lights, a 75 volt bulb is now available which eliminates the need for the resistors. This is not unusual in the case of the single bulb Gyralites which have a 74 volt motor (17550, etc.). This recent GE 75 volt PAR-56 bulb is rated as 350 watts. The conventional PAR-56 bulbs are 200 watts @ 30 volts. It should be noted that resistors were dispensed in values which would conform to the use of the Gyralite. The resistors to accommodate dimming of the Gyralite bulbs were also included in the mounting enclosures along with the line dropping resistor.

In a single Gyralite using a 30 volt bulb, a resistor(s) would also have to be used. If the red bulb was not going to be used in a dual Gyralite, the 2 bulbs were wired in series. This applied to the dual Mars Light units as well. (this was mentioned above)

The use of resistors in a series dual bulb Gyralite or Mars Light is "hit or miss". Pyle- National made specific adjustable resistor(s) for this purpose. I have corresponded with sources who claim that they don't bother with using the resistors. They claim the bulbs can take the 36- 37 volts without using the resistors. This applies to the wiring up of 30 volt constant ditch lights as well.

One source claimed the PAR200 bulbs maximum voltage is 40 volts. There are formulas to figure the bulb life based on exceeding a bulb's rated voltage. The problem is that GE and Osram Sylvania are sticking to 30 volts being the maximum design voltage, whereas others say it is up to 40 volts.



The electrical schematics show the Gyralite or Mars oscillating light being supplied off a 30 amp breaker. The motors have a continuous current draw of 1 amp or less (12 volt Mars motor: 3 amps). One can see that anything that jams the motor will cause it to burn out. Apparently railroads did not fuse the motor to prevent this because they would rather replace the motor from an all out failure than to figure out why a fuse blew due to perhaps ice or other condition that might not show up as the cause.



Locomotives through the 70's were built with one circuit breaker for both headlights (front and rear.) This changed to a breaker for the front and one for the rear. Again so the minimum is affected if one circuit fails on the road.


As for the Gyralite breaker values, the SP's B30-7's had full sets of lights on both ends. These circuits were protected by a 30 A breaker. The D&H U33C's with only one light also used the 30A breaker. Remember too, the SP units had a transfer switch so only one light set could be on at one time. Therefore, it looks like the 30A breaker was typically the size used.

I would not worry about individual fuse protection of the motor etc. It is the same in your house. You have 15 and 20 A circuits for lights etc, yet individually they only draw 1 or 2 amps. I think the designers know that if the motor is going to fail, it will be replaced anyway. So if it "burns up" a fuse really won't make a difference. If they did fuse it and the fuse protects the motor from burning up a few months before it actually fails completely, then the locomotive will end up in the shop twice. So did the fuse really save anything? They probably spent more money in shop time and non revenue time than the cost of the motor (which will be replaced anyway.)

A real life example of this is the original Santa Fe Alco PA's. They were originally designed with all sorts of protective circuits designed to protect individual circuits as you propose. Unfortunately, they caused so many false failures etc. that all of the extra stuff was removed to keep the units out of the shop. After the modifications, they performed much better.

Another point: If everything is protected with a fuse etc., it will take longer to determine why say, the Gyralite stopped turning. If the fuse blew, they would have to determine "why" did the fuse blow. Maybe there was some ice on the shaft when it first started so it drew heavier current to start. But the shop personnel would not know that and have to take the time to figure it out. On the other hand, if there are no fuses, the motor will continue to function until it literally blows up. When it arrives at the shop, it will be obvious to determine what is wrong. Remember too that the shop personnel are not trained in all areas. The prime objective is to get the unit in service as fast as possible and keep it out of the shops as long as possible. Therefore, the fewer the parts to break, the better.

Bill Kaufman