How do you prove an LED is actually lit?

[MarkS]

12v vehicle LEDs are known to get very hot. A simple thermistor might suffice. If it drops below normal operating temps while on, something is wrong.


[edit] Helps to read the post before posing...

[Infraviolet]

The OP seems to be in the UK, unless he's in northern Scotland snow is pretty rare here. And I guess the heritage railway would be shut on days if the weather is really bad anyway.


The picture of what he is replacing is pretty bulky, so he looks to have plenty of physical volume to work within. In his position I'd look for high powered LEDs/arrays which do NOT have internal driver circuitry, then PWM them very fast and monitor a light sensor for this varying signal. The trickiest part will be finding an LED type which is bright, doesn't have internal driver circuitry (which would prevent PWMing), and can be easily driven from a simple fixed DC voltage level without him needing to build a complex external driver circuit (the way one can need to for some fancy LEDs, especially lasers but it can be the case for simpler high power LEDs too). In the volume of that box he'll also need to work out a way to provide a regulated power source for the logic and op amp chips involved in monitoring the light sensor, he doesn't seem to have said exactly what power source is currently available (only that the LEDs he has tried can take 12V DC or AC).

[Jeff eelcr]

The LED light assemblies we use here are not all the same they have double sided and different size ones.
Some gates are longer with more lights, all lights are DOT spec'd.
For the flasher units + (power supply) we had to pass railroad testing including artic tests.
Jeff

[IDEngineer]

I'm thinking of two LED arrays, toggled in operation, where during one phase the first array is the emitter and the second array is the detector. During the other phase their roles reverse. Both arrays point in the same direction and detection relies on some internal reflection within the fixture. If you optically associated LED's from opposing phases you could know which devices had reduced output. The advantage here is efficiency... all devices contribute to both emission and detection.

[peter-h]

An LED, when not lit, is also a photodiode.

In some designs, this is used to measure the ambient light and adjust the ON current accordingly.

So you could have a piece of photoluminescent material next to the LED which returns the LED light back to it, with the material selected to have a short TC, say 1 second, and you periodically turn off the LED and measure how much light is coming in.

One problem would be how to exclude the effect of ambient light, but that should be possible because you will know the rough TC of the luminescent material.

[mikeselectricstuff]

I have seen matrix roadsigns with a photosensor pointing at every LED.
A phototransistor will probably do if you can get decent optioncal coupling, so no amplifier needed.
If you don't need continuous monitoring, you could connect multiple sensors in parallel and periodically light each LED in turn & look for chnages.

[BrokenYugo]

The suggestion to modulate the LED and detect that seems best to me, the matter that we're discussing DIY modification of safety equipment aside.

That said, you mention there are optics in these lamps. It's been my experience, but by no means a hard and fast rule, that you get bad results retrofitting any lamp that's more complex than a simple reflector. I assume because they're designed to focus on a white hot filament in a fairly precise location, not an array of SMD LEDs. For example there's no real upgrade to halogen car headlights (aside from cleaning fogged housings and minimizing voltage drop in the wiring), whatever else you stick in there inevitably screws up the cutoff, giving the impression of of more light, while blinding oncoming traffic.

[Kleinstein]

Just in case it's remotely of interest this is one of the lamps taken out of it's housing.

It is of course of enormous interest, since you have started a thread asking about this problem, but you have been strangely starving us of information, making everyone guess and speculate, as you can see above.

Is it perhaps the case the bulbs you are trying to replace are similar to this one?

https://www.proflamps.com/us/en/Products/Applications/Airfield-%26-Railway/p/13800306842296/?currency=GBP

Given that the life expectancy is only 600 hours, I can see why replacing them could become annoying. On the other hand, given that the indicated price is £1.45 per bulb, and each bulb will have to be replaced every 25 days, you will need something very inexpensive as a replacement to provide a reasonable payback on the investment, I would think?

It is not just the costs for the new lamps (which seem to be higher by a factor of 10 - though likely with a discount for larger quanteties). Often much of the costs are for sending out a maintainance team to sometimes relatively remore locations. At least less ladders needed for the rail signals.

When introducing the LED trafic lights the longer service intervals were a big plus, not so much the energy savings.
Depending on the environment one may still need regular cleaning (e.g. check for spider webs).

Modern railways have an electronic signal going to the train as a backup. The optical signal is limited reliability becaus of the human factor anyway, no matter how good the lamps.

Normally I would consider lamps with  3 redundant strings and than a somewhat more sensitive system to detect a drop in the current good enough. 2/3 the intensity should be good enough until fixed.
A failure with reduced optical power when getting the right current is rare and mainly an issue when running too hot or a series failure with white LED loosing the phosphor. Even if 1 LED in the string would fail, this would be something like 1 in 3 or 4.

Modulation of the lamps would be a way to make it easier for an electgronic system to detect them too - but that would be a larger change to the system.

Chances are this would need changes in the regulations anyway, no matter how small the change.

[TomS_]

Maybe it would be going further than is technically possible or feasible (or even wanted), but if you were inclined to produce your own replacement "bulbs", there are LED drivers available that can detect a variety of failure modes (like opens and shorts), which could then be read back using a small microcontroller and perhaps operate a set of relay contacts that can indiciate a fault to the rest of the signalling system.

For example, https://www.ti.com/lit/ds/symlink/tlc5916.pdf

I used some of these in a project to drive 16 segment LED displays, although I didnt use the error detection functionality because my shift register doesnt have an ability to read back, but its there.

Maybe there are smaller chips available, or maybe these could be built onto remote boards that then drive the LEDs so they dont have to try to be wedged into a tiny space.

Just a thought.