Why do US trains have flashing lights on the front of them?

May 29, 2004 4:45 PM

Why do US trains have alternating flashing lights on the front of them? [Er, that's it, I'm just curious]
posted by bonaldi to travel & transportation (23 comments total)

They're called ditch lights.

Quoted from here:

To increase train conspicuity, many railroads have equipped their locomotives with external auxiliary alerting devices such as strobe lights, ditch lights, crossing lights, oscillating devices, paint schemes, and reflective materials.
posted by tss at 4:50 PM on May 29, 2004

conspicuity

In the following week, I will try to work this word into my conversations whenever possible.
posted by SPrintF at 5:42 PM on May 29, 2004

Sub-question: How is conspicuity different from conspicuousness? They both seem to be noun forms of conspicuous. I've always used conspicuousness, ignorant as I was to the existence of conspicuity.
posted by ChasFile at 6:32 PM on May 29, 2004

Further sub-question: Why do pairs of flashing warning lights always seem to have different periods? I first noticed this on the lights of the George Washington Bridge, but I think the phenomenon reaches its most dramatic with the two red lights atop the Washington Monument, which are set very close together. One light blinks, say, once every two seconds and the other blinks once every 3 seconds. The result, of course, is that they blink together, and then slightly offset, then slightly more, until they blink in perfect one-on one-off, and then slowly begin to synchronize again, until they blink together at the same time, and the whole thing begins again. I've always found it oddly mesmerizing.

The point is, is this technique merely a result of technical considerations, like each light being on circuits of slightly different voltage or current? I've always fancied that they'd be on different curcuits, so that if one blew the other would still go, and figured that might have something to do with it. Or are blinking lights of differing periods more eye catching to the wayward helicopter pilot?
posted by ChasFile at 6:40 PM on May 29, 2004

Chas - I think it's more like "They didn't care when they set it up if they blink at the same time"
posted by SpecialK at 6:57 PM on May 29, 2004

I refuse to accept that. Clearly I've put way too much thought into this already, and I require a satisfactorily complicated anser. Please just indulge me.
posted by ChasFile at 7:17 PM on May 29, 2004

i'm with ChasFile. Even if they didn't care, it'd be more likely that the lights would be the same period, that is with a constant offset. Why would the periods be different? I'd like to know this too.
posted by vacapinta at 7:38 PM on May 29, 2004

Coo, thanks tss.

ChasFile: It's because they're on different circuits, in case one blows, you know.
I made that up. Don't know where I got it from
posted by bonaldi at 7:43 PM on May 29, 2004

From a stand-point of function, the differing periods make them even more difficult to ignore... our brains are wired to find patterns.

You may have noticed how, about 5 or 10 years ago, emergency vehicles across the nation started sporting asynchronous strobe lights... same reason.
posted by silusGROK at 8:05 PM on May 29, 2004

I agree with silus. If they are in harmony, then you can become mezmorized by the blinking. Plus, as silly as it sounds, it is easier to notice two lights with two flashing different cycles than two different lights with identical cycles.
posted by jmd82 at 8:09 PM on May 29, 2004

Here are the standards (PDFs):

Advisory Circular AC 70/7460-1K : Obstruction Marking and Lighting
150/5345-43E Specification for Obstruction Lighting Equipment

The 1st one says that lights on the same structure should be synchronized. "The red obstruction lighting system is composed of flashing omnidirectional beacons (L-864) and/or steady burning (L-810) lights. When one or more levels is comprised of flashing beacon lighting, the lights should flash simultaneously." The Washington Monument might just be wired wrong or they don't really follow the standard in reality.

If they are on different structures, that's another story. The flashing rate has a 5% tolerance so they could easily be slightly out of phase. Also, some light
posted by smackfu at 9:05 PM on May 29, 2004

While we're on the topic: the CN Tower's lights are set up strangely. It has two colours: white and red. There are four red lights evenly spaced from the ground to the main pod. Then there are four white lights evenly spaced from the pod to the top. But then there are four more red lights, all bunched together at the top, one for each width of the tapering tower. The white lights all flash together, very quickly. The red lights all flash together—both the ones on the top and the ones on the bottom, in synch—a little more slowly. Anybody have any clue as to the regulations behind this setup?
posted by DrJohnEvans at 9:28 PM on May 29, 2004

Putting on my EE hat for a moment:

If you're an engineer trying to build a cheap yet highly reliable circuit that flashes a light, then you might generate time delays by waiting until a capacitor charges to a specific voltage threshold. It's cheap, easy, and robust, but the accuracy is limited by the error of the capacitance (+/- 15% is not uncommon). So you might choose components with nominal values that should give you a two-second flashing period, but in reality you end up with some distribution of periods. Odds are good that the periods of any two flashers will differ noticeably.

If you want really accurate timing, you might put together a digital circuit of some sort. That would be both more expensive and easier to screw up, and it's complete overkill when all you want to do is flash a friggin' light.

This is just an educated guess, of course.
posted by Galvatron at 9:30 PM on May 29, 2004

Galvatron: yes, but it takes a 5¢ preset (the little screw-driver tunable variable resistors) to be able to set them to the same frequency.

I think an added bonus of having them be of different frequencies is that the period they're both off is shorter, so they'll be noticed a few centiseconds earlier in the worst case. You could of course put them at the same frequency in opposite phase, but that'd look like an always-on light from a distance, and as was already pointed out, the brain sits up and takes notice of things out of sync.
posted by fvw at 9:53 PM on May 29, 2004

Galvatron: yes, but it takes a 5¢ preset (the little screw-driver tunable variable resistors) to be able to set them to the same frequency.

The same frequency? To what tolerance? In analog electronics, "equal" never really is. You can always measure to one more significant figure, and eventually, there will be a difference.

It's like expressing 1/7 or pi in decimal form. You can get really, really close, but never exact.

To affirm the EE that posted earlier: it's component tolerance. Caps are commonly +/- 20%, resistors 5% or 1%. Of course it's possible to design circuits that depend less on component tolerances, but in the end it's probably not worth the extra cost...
posted by lalas at 10:02 PM on May 29, 2004

it seems like every light i've seen like this has been synchronized.
posted by rhyax at 2:20 AM on May 30, 2004

The lights atop the radio towers at the Portland Skyline cemetary are like this. I remember walking in the night with friends, gently being persuaded into a benevolent mindset of capriciousness at the duality of headstone and electricity in the summer night, the amber blinks of the towers above speaking of late nights of youth and cheap wine. Nary a moment without a girl by my side, we would saunter off into the blackness, drunk and in love.
posted by Keyser Soze at 2:44 AM on May 30, 2004

even if they were digital they'd have to share the same clock. synchronized lights are probably all on a single circuit, controlled by a single switch.
posted by andrew cooke at 5:15 AM on May 30, 2004

Better to have them wildly unsynchronized than minutely unsynchronized.

If they were tuned by hand, f'rinstance, the tolerance couldn't be better than, say, 1/10th of a second over a hundred flashes. Getting that accurate would require microscopic adjustments and timing out the results over a good half-hour or more.

Which, in the end, would be pointless, because some time later they're going to be completely out-of-sync, and then will sloooooowly come back into sync.

Might as well forgo that, and let them go in and out of sync over the space of a minute.
posted by five fresh fish at 10:27 AM on May 30, 2004

I believe the FAA regulation mentioned above refers to multiple lights in the same structure at different levels, for example in an antenna tower or a smokestack. Having the lights flash simultaneously helps to visualize these narrow structures. From what I remember in antennas and such the lights flash in sync, sometimes even across different smokestacks or antennas in the same place.
posted by golo at 12:19 PM on May 30, 2004

The deframbulator which regulates the lights on the Washington Monument is precisely calibrated to de-phase the blinkages of the warning lights to maximify the prophylactic effect. The human brain decodes deframbulated signals at an accelerated rate.

Now can we go back to talking about conspicuity? It sounds presidentialified to me.
posted by theora55 at 1:02 PM on May 30, 2004

Conspicuity is big-time in the motorcycle rider world. It's the name of the game: get seen, so you don't get hit. Hence the retroreflective patches, and hence my preference for bright yellow jackets, white helmets, and blinking headlights/taillights.

There are those that claim "get heard," foolishly thinking their loud pipes can be heard inside soccer mom's minivan when she's got three screaming kids, a cellphone to her ear, and the radio on.

Whichever way floats your boat, though: between conspicuity and awareness, one hopes to avoid getting hit by the seemingly endless supply of braindead cagers out there.
posted by five fresh fish at 2:06 PM on May 30, 2004

There's a post by JWZ about <blink/> and how they changed from on-off to on-on-on-off due to epileptic seisures.

(I'm cracking up laughing now thinking of a doomed epileptic spasming on train tracks as the on-off lights an approaching train speed closer - oh jesus that's funny - he's going to die - oh my ha!)
posted by holloway at 4:47 PM on May 30, 2004