Semi OT: LED traffic light failures

If there are clusters then it must be strings of in series connected LEDs. If it's flickering, then it must be the LED-controller or -driver. If it's only green... maybe green LEDs require more current for the same amount of light, hence blow up their drivers earlier?

Just some uneducated thoughts...

joe

I count at least 67 manufacturers and distributors: or more:

What's the name of this company? I'm curious and may want to buy some stock in this monopoly.

On what planet? In the distant past, I had something to do with installing Metricom radios on lighting poles. Everything was 117VAC. However, the wires were labeled 600VAC maximum insulation, so that may be where you found that number.

However, the street lights vary substantially. Common voltages and phases are: 120/240V 1P and 3P 120/208V 1P and 3P 240/480V 1P 277/480V 3P To the best of my limited knowledge (and some Googling), 600V is not used anywhere.

Not withstanding a few minor efficiencies, all LEDs give light relative to the current. One electron equals one photon. Green and blue LEDs require a higher drive voltage than red or yellow. There would likely be fewer green LEDs in a string than red using the same driver circuit. Each green LED uses about 40-50% more power but I don't know how that relates to the heat generated since a lot of the power is emitted as photons.

My daughter has a set of multi-color Christmas lights hanging in her room (for ambience) and most of the greens in the string are dead, while very few of the red, yellow or blue are.

On Tue, 28 Jul 2015 18:27:16 -0700, Jeff Liebermann Gave us:

No. The HPS lamps do not run on 120VAC. It has nothing to do with the rating of a standard THHN line feed.

Traffic lights do not use HPS (high pressure sodium) lights. Street lights do use HPS lamps. If you're unable to make the distinction, the HPS lights are kinda yellowish and are useful for seeing where you're walking or driving. Traffic lights are red, yellow, and green, and regulate the flow of traffic.

"Glenn Newman tried to pry Williams loose from the metal grate, but the estimated 5,000-volt current was too great."

Current in volts? 5,000 volts in a bus stop? Held glued to the gate because of some mysterious attraction? Incomplete article without a subscription? You can do better than that. Please find a better source for traffic light power supply voltage(s).

On Tue, 28 Jul 2015 20:13:02 -0700, Jeff Liebermann Gave us:

If you read my original response post, illiterate idiot, you would note that I was and still am referring to street lighting. DOH!

Sorry but I failed to notice that you changed the topic in the middle of the thread. That would be this posting: The message to which you were replying was about traffic lights, as were the other articles in the thread. If you're going to change the topic, it would be helpful if you would better identify the transition.

Can I assume that "signal maker" means traffic signal when you posted: "There are not that many signal makers in the US. All 50 states use the same asshole company who STILL uses the same broke dick day/night sensor on street lamps, despite RTCs being around for decades now."

Could I trouble you for the name of this company that you claim has a monopoly in the allegedly limited "signal maker" market? I want to invest.

Also, none of this changes the not so trivial factoid that no traffic light or street lighting system uses 600 VAC for power.

On Tue, 28 Jul 2015 21:25:14 -0700, Jeff Liebermann Gave us:

I also spoke about traffic signals (only laymen call them traffic lights) in that same post, and it was my first post so it wasn't 'in the middle' of anything.

On Tue, 28 Jul 2015 15:49:10 -0400, Spehro Pefhany Gave us:

LEDs are not called "bulbs".

As to the green fails more than the others, I think they were lame to use colored LEDs to start with as the signals that were illuminated by incandescent 'bulbs' were given their 'color' via the 'lens' they were behind.

They should have used white throughout, and continued to use lenses to provide the tinting.

I don't think LED's are that good. (Well at least the few I looked at... mostly high power through hole's.) You have to ignore all the luminous efficacy stuff...

George H.

That would be extremely inefficient. "White" LEDs emit mostly blue and yellow, hardly any red.

The blue, the main output, would be wasted by r/g/y filters.

Green LEDs are very efficient.

I've seen LEDs fail in a blinking mode. I don't know what the mechanism is, but it might be thermal wirebond/bonding pad problems. They do seem to fail open, not shorted. I've heard that greens tend to fail more in this mode than red or yellow. I've personally seen it happen in white LEDs.

Those are white LEDs used for lighting. They generate a monochrome light which shines on phosphors to generate other wavelengths of light to make them look white. So there is a loss of efficiency there. I think these numbers are also inclusive of all losses including power supply and absorbed light in the assembly.

But it looks like you are right. I see even plain red LEDs are only in the 50% ballpark today. Here is a wiki link to a color based efficiency table. I don't see how they are calculating the W/W efficiency. Comparing the red and red-orange LEDs, the efficacy goes from 72 to 98 lm/W while the efficiency goes from 0.39 to 0.29 W/W. I thought red and red-orange LEDs had about the same voltage drop, so this doesn't seem right.

Here is an interesting report I ran across...

On Wed, 29 Jul 2015 09:17:16 -0700, John Larkin Gave us:

What part of LENS do you not understand? Oh and it would have been more 'efficient' than the previous incandescents that were in use, and the reduced wattage requisite makes the entire premise of 'efficiency' moot. as they used tinted lenses as well. They would have been fine and far lower cost as the industry spent all their time making white more efficient, and more power, so an array of white LEDs of all 1 W each yet powered at half rate would last twice as long and operate half as warm. Replacing a single color would be cheaper and easier to stock as well.

You lose.

Please accept my humble apologies for not recognizing your oscillatory change of topic. However, it still would be helpful it made the transition a little more obvious.

Now that that's taken care of, could you kindly supply the name of the company that you claim has a monopoly in the allegedly limited "signal maker", "traffic signal", or "traffic light" market? I'm really interested in investing in what your description suggests might be a new and undiscovered monopoly.

Incidentally, I really am a layman when it comes to traffic signals and traffic lights. My only experience is inspecting the remains after someone crashed into the lamp/light/traffic/signal/sign pole. As I vaguely recall, the insulation was more in line with 120/240v service than 600V. Any information you might have on street lighting that operates on 600V would be interesting.

Anything you do to filter white to achieve the color you want will incur losses. Lenses are expensive even in plastic. Measurements have been made and colored LEDs win. You can educate yourself with the MUTCD found here:

Or not.

I have in front of me a cheap torch with 9 white 5mm LEDs. They are connected in parallel and powered from three AAA cells. There is no current limiting or other circuitry.

At the moment, three of them flash, one is off and five are steady. I've taken the PCB out of the torch, and if I bend the LEDs about I can sometimes make one of the flashing ones go out.

Oh look, there are now four flashing. At one time, I thought they were all flashing in phase, but it's clear now that they are all flashing at different frequencies, maybe between 5 and 15 Hz, and at different mark/space ratios.

The PCB solder joints get quite hot (ouch!) for the LEDs which are continuous.

When the torch was bought, batteries were included from some obscure manufacturer. These have since been replaced with Duracells.

What I think has happened is that the Duracells have a lower internal resistance than the original cells, and are supplying much more power. This has caused the bond wires to break in such a way that they reconnect when cool and disconnect when hot.

So my carefully conducted and extensive accelerated stress testing has resolved the great traffic light problem.

Cheers

Did you apply summertime heat in an enclosure?