InterlockingTower, in the interest of respectful discourse, your post would have been considerably improved if you had only posted the second paragraph.

I think that is arguable. It depends on what your organization's mission is. If you feel that preserving technological history is important to your history, then maybe substituting LEDs is the wrong answer.

Taken to an extreme, the thought that "it does the same thing, so get over it" could justify some stuff that you as (I guess) a steam loco fan would find offensive. I mean, the whole boiling water traction apparatus is very difficult and expensive to maintain, subjects you to all these ornery FRA regs. So, why not replace all that with a diesel? It does the same thing -- move the train -- right?

Now on the other hand, something like a cab light, it could be argued, is just an inconsequential detail and replacing an incandescent with an LED does not affect the appearance or the style of illumination in any way that can be seen.

Changes like light bulb type are probably inevitable as time goes on. Other examples would be using modern polyurethane paints instead of oil-based paints, using modern pipe sealants instead of white lead, etc..

Such changes should be documented in the files of restoration of the artifact.

Two things to be aware of with LED bulbs:
1) they are dependent on proper polarity! Yes, even in an AC environment. While resurrecting the lighting in a demotored MP54 in preparation for a movie gig, I was reminded of this. We were not only reactivating the long-disused lighting, but were running them off a small genset, so the minimal wattage of LED's versus incandescent bulbs were a no-brainer. However, the cabling guys in the 1930's weren't concerned with polarity on the light circuits, and several banks has flipped connections. Of course, those banks wouldn't work until (as it were) the light went on in my head, and I flipped the leads.

2) If you're doing a movie gig, cinematographers do not like the Kelvin light ranges of LED's, regardless whether they're warm, cool, or whatever. After all that work, the bulbs were replaced temporarily with incandescents for the shoot.

You must have been unlucky. LEDs as electrical components are certainly polarised, but the LED bulbs we use don't care which way round you insert them (ours have bayonet fittings, not edison screw bases). I tested them to make sure, I thought it might be a problem too. Maybe ours have a bridge rectifier in them to make them polarity agnostic.

AC power makes them flicker noticeably. Our 50Hz mains may make this worse.

Yeah, these were consumer-grade dimmable GE edison-screw base bulbs. Non-dimmables are notorious for strobing -- and the trick on those is to add a load on the circuit -- an incandescent bulb.

That's a slightly different case. When the original material becomes "unobtainium" because we're afraid of lead, or VOCs, then there is no choice. I doubt the production of incandescent bulbs will ever be outlawed entirely. There will still be specialty shops that can make them, just as people are still making ammo for guns that haven't been made in centuries. But these bulbs might be very expensive, vs very cheap LED conversions. It depends on your mission statement and how you interpret it.

For the traction preservation world, one of the challenges is getting a screw-in LED that is going to look 100% authentic (very often the bulbs were exposed and the little plastic thing at the base where the electronics unit is housed is an obvious anachronism) and also respond the same way. That means having a "warm" color temperature, and also dimming and becoming redder when the line voltage drops.

The other issue is the 5-in-series wiring. Don't trust your home despot off-the-shelf $4 LED bulbs to work correctly on a 600V system where 5 of them are wired in series. The driver circuitry is looking for a ground-referenced voltage input. If you think about the input circuitry, normally you'd have a rectifier followed by a filter capacitor, and then a buck converter running in constant current mode to drive the LED or LEDs inside. If all the bulbs have identical capacitors then on startup the voltages can divide equally and the thing could work. Now substitute one bulb from a different manufacturer and all bets are off. If during startup the voltage that develops across some bulbs is say 50V and others develop over 200V, are the components inside rated to tolerate that voltage, even briefly? What happens when lightning strikes the trolley wire?

Aren't thoes bulbs also looking fo AC and not DC.

An AC LED bulb should be fine on DC if they’ve incorporated a bridge rectifier. I’ve never opened one to find out.

Good point about wiring them in series, though. Methinks you’re gonna have regulated voltage looking to go to ground from a bridge rectifier or switching power supply, not to straight another regulated supply. It may not work.

I think there are multiple issues to consider.

- does the LED bulb look exactly the same, visually, as the historically correct bulb?- Purchasing amber instead of white bulbs, and making sure the housing hides the fact that they are LED can go a long way.

- is this a museum? Or a tourist railroad.

The bulbs I posted a photo of do have a rectifier and tiny ballast. Polarity not an issue. If I spray “frost paint” on them they look identical to the old stock 34 volt GE incandescents. The LEDs are 2700K, but the GE bulbs are less bright and a little warmer in color.
That said the object here is to allow visitors to see whats in the car through the windows. Many times with half the 25 watt incandescents burnt out it was like peering into a cave. Also with all the visitors outside, and tours I’ve taken through the car inside, I have never overheard a word about the accuracy of the light bulbs. In addition I’m not so sure an accurate 1920 era bulb is available, anywhere?
We did have a volunteer go to great expense to get “real” carbon filament, Edison bulbs made for an old Chicago L car. They put out less light than a couple watt night light. As someone mentioned they are 5 in series for 600 volts. The first time the pole jumped through a frog and arced, every one of them blew out. There is probably another thread here that goes over the extra things a special traction bulb needs to keep it from frying things when it opens and has 600 volts across the base.

Tell me if I’m wrong but if you have five LED bulbs, or incandescent for that matter, in series, wouldn’t you have the cumulative current draw of the following bulbs through the preceding bulb? For example, the first bulb in a string of five is going to have to be able to pass the current drawn by the four after it. Are the teeny tiny regulators in those LEDs robust enough to handle 5X their design current the and last?

Would you’d have better luck in the long run stepping down the voltage to 120V or less with self resetting over-voltage protection or something to temporarily shunt the voltage spikes and then wiring the bulbs in parallel?

I don’t know much about traction and that solution certainly wouldn’t be period prototypical but could be a suitable solution to blowing out the bulbs every time the pole arcs.

Dunno. Just kinda thinking’ out loud. Hopefully someone smarter than me would chime in.

To operate five LEDs in series from 600 volts couldn't you parallel each bulb with a resistor? Resistor values would be selected so that most of the current flows through the resister. This negates a lot of the efficiency benefits of LEDs but at least you could use the readily available bulb choice. I don't have any suggestions on how to make the bulbs dim and change color when voltage drops, though.

That's not how electricity works.

Uhhhh…. I think it does. Ya know, like, Ohm’s Law and stuff.