NYC's 250,000 Street Lights To Be Replaced With LEDs By 2017

An anonymous reader writes "New York City Mayor Michael Bloomberg announced that the city's 250,000 street light fixtures, which currently use incandescent bulbs, will be replaced with LEDs by 2017. It's part of a plan to reduce the city government's emissions by 30%. The LEDs have a lifespan of 20 years, more than three times that of the current incandescent bulbs, and Bloomberg says it will save $6 million in energy and $8 million in maintenance every year. It will be the largest LED retrofit in the country. 'The first of three phases to replace the standard "cobra-head" high-pressure sodium street lights, which will upgrade 80,000 at a time across the five boroughs, is expected to be completed in December 2015 with the final phase expected to be completed by 2017. Following the replacement of roadway lighting, decorative fixtures in the city's business and commercial districts will be addressed.'"

Comments

I wish they'd do it here.

[mcgrew]

I wonder how many smaller cities have already done this?

Re:I wish they'd do it here.

[madirad]

We've got it here in Oakland. The lights were changed about 2 weeks ago, at least on my street.

Re:I wish they'd do it here.

[AmiMoJo]

We have had them in my crappy city in the UK for a couple of years now. They put out better light than the old orange bulbs and seem brighter. The orange colour of the old bulbs is actually known to inhibit night vision, so white LEDs are safer.

Re:I wish they'd do it here.

[EvilSS]

I think you're confusing street lights and traffic signals. Places in the snow belt have had issues with LED traffic signals getting blocked with snow, but I can't see the same thing happening with a downward facing street light.

Re:I wish they'd do it here.

[mcgrew]

Yes, we've had LED traffic signals here for years, and I've only seen them obstructed by snow once. You need a wet, sticky snow and a swift drop in temperature for it to happen. IINM they put remote-controlled heaters in the newer ones.

And it seldom snows upwards. I don't think I've ever seen it snow upwards.

Re:I wish they'd do it here.

[Animats]

I wonder how many smaller cities have already done this?

Redwood City, CA, near where I am, is doing it. It's striking, because Redwood City standardized on yellow sodium lamps some time in the 1930s. You know you're in Redwood City when the street lights turn yellow. The new daylight LEDs are a big improvement.

If your community is doing this, push for solar power on some of the lights. Not necessarily all of them, but at least at street corners. That way, no matter what disaster happens, some lights will stay on.

Re:I wish they'd do it here.

[drnb]

I think you're confusing street lights and traffic signals. Places in the snow belt have had issues with LED traffic signals getting blocked with snow, but I can't see the same thing happening with a downward facing street light.

It will when temperatures fluctuate above and below freezing. When above the snow on top will melt and the water will cling to the downward facing cover where it will freeze when the temperature drops. Icicles may form, which could be a safety hazard.

Re:I wish they'd do it here.

[drnb]

This is a pretty terrible idea. It snows in NYC right? the LED's dont generate enough heat to melt the snow, so they will have to install heaters to keep the street lights from caking up with snow/ice and becoming useless.

Even if heaters need to be installed they only need to be turned on during cold weather so there is still a potential efficiency benefit.

Re:I wish they'd do it here.

[GonzoPhysicist]

Monterey, CA did it a little while back. One nice side effect I've noticed is less light pollution compared to the presidio which didn't change their lights. No idea about the cost savings though.

Re:I wish they'd do it here.

[Bob the Super Hamste]

There is even a simple solution to that as well. We have lots of traffic lights with LEDs in them here in Minnesota now and they all seem to work fine in our long cold snowy winters. They make devices that almost completely surround the light to shield it from snow and they work great and even the non LED traffic signals have them (as long as I can remember traffic lights have had the shrouds around them here).

Re:I wish they'd do it here.

[ebno-10db]

Part of my town (the main streets) have LED street lights, and we have no problem with snow or ice. BTW, LED's are in the neighborhood of 20-30% efficient, so they don't run ice cold (pun intended). It may seem so though if you've only touched indicator LED's (flashing lights on equipment).

Re:I wish they'd do it here.

[Anonymous Coward]

If your community is doing this, push for solar power on some of the lights. Not necessarily all of them, but at least at street corners. That way, no matter what disaster happens, some lights will stay on.

Assuming you add battery replacement and panel replacement costs to your maintenance budget.

Re:I wish they'd do it here.

[ebno-10db]

I don't think I've ever seen it snow upwards.

Never seen a good blizzard?

Re:I wish they'd do it here.

[Anonymous Coward]

And it seldom snows upwards. I don't think I've ever seen it snow upwards.

When it finally does snow upwards here in NYC, the last thing I'll be worried about are the traffic signals.

Re:I wish they'd do it here.

[fuzzyfuzzyfungus]

I wonder how many smaller cities have already done this?

I think that it's not uncommon (though traffic signals usually go first, since LEDs have been cheap and good at red, green, and amber for longer than they've been either cheap or good for white, and bulbs-behind-filters have always had even more miserable efficiency than bulbs in general).

LEDs are still pretty expensive, and white ones (because they are usually blue ones pumping a phosphor layer) are still less efficient than one might like; but one big advantage is lifespan.

A replacement lightbulb doesn't cost much; but sending out guys in bucket trucks to deal with dead ones adds up.

Re:I wish they'd do it here.

[K. S. Kyosuke]

Well, white LEDs are most likely lousy for low levels of illumination because our color perception shifts with the illumination: our visual cortex expects redder colors in darkness, so physically white faint light looks unnaturally blue, and incandescent light bulbs correspondingly look too red when you attempt to use them for daylight levels of illumination. I'm patriotically proud to point out that this is called the Purkinje effect. ;-) While the fact that LED light appears brighter may lead to energy savings beyond the simple increase in energy conversion efficiency, I wonder how it will change the perception of traffic signs. The red ones will probably appear even darker. What about traffic safety?

Re:I wish they'd do it here.

[ducomputergeek]

I've seen it snow upwards before. I was in a city and on the 7th floor of a building. Apparently they got updrafts when the wind was blowing from the north and it was indeed snowing upwards!

Re:I wish they'd do it here.

[Joce640k]

This is a pretty terrible idea. It snows in NYC right? the LED's dont generate enough heat to melt the snow

Yes they do.

Re:I wish they'd do it here.

[fahrbot-bot]

The orange colour of the old bulbs is actually known to inhibit night vision, so white LEDs are safer.

My understanding is that white light destroys (photobleaches) the rhodopsin/visual purple in the eye, inhibiting night vision, which is why ships and subs use red (or simply non-white) lights at night. Anyone out there with more specific info and/or why one color is better than another for night vision preservation?

Re:I wish they'd do it here.

[Joce640k]

Will this stop the streetlights going off when 'psychic' people walk under them?

Re:I wish they'd do it here.

[Anonymous Coward]

I have read that cooler colour temperatures in the whitish/bluish range actually helps to inhibit sleep. That could be a good thing to keep drivers from dozing off.

Re:I wish they'd do it here.

[SJHillman]

We have shrouds on many of our lights in upstate and western NY, although they're far from ubiquitous. I've never really noticed a big problem with snow sticking to traffic lights except in the rare blizzard that combines the right amount of gusty with large, wet snowflakes and relatively warm temperatures... it's a once every few years occurrence. It's far, far more common for traffic signals to lose power altogether.

Re:I wish they'd do it here.

[Anonymous Coward]

I live in Sweden, plenty of snow and plenty of LED traffic lights here. Have never seen one covered in snow but I guess it happens in certain conditions.
However, what to do when you encounter a broken traffic light is taught when learning to drive. (Assume the worst, be as cautious as you can be, fall back to regular right-of-way rules if possible, etc). If you get yourself or someone else killed because you couldn't see the colour of your light, its your own god damn fault.

Re: I wish they'd do it here.

[FishTankX]

You could largely eliminate battery maintenence by adding nickle iron batteries which effictevely have infinite endurance. There are some nickle iron batteries which still function 100 years later.

Re:I wish they'd do it here.

[Minwee]

When it snows hard enough to cover the _bottom_ of the streetlights, where the light comes out, then you've got bigger problems than that.

Re:I wish they'd do it here.

[adolf]

I've seen regular incandescent traffic lights clogged by snow, too. Especially the yellow light, or whichever red/green light defaults to "off."

It's just part of driving: If you can't see what the traffic light is indicating (be it because of dead bulbs or power outages or snow or whatever), stop, look out the window(s), and see what's happening. Go when safe.

NBD.

(That said, this discussion happens whenever anyone mentions anything about using anything LED in a public space. The answer to the snow problem is and has always been heaters, and it's the same answer no matter if it is LED or incandescent or carbon-arc or...)

Re:I wish they'd do it here.

[digitig]

They've started introducing LED street lights in my UK city too, and the light they give is far better than the old sodium lights. Even if the LEDs don't last as long as claimed, I understand they only have to last a couple of years for them to have paid back the higher cost of the luminaries.

Re:I wish they'd do it here.

[Anonymous Coward]

I've been in a couple of good blizzards in England. Believe me, it *can* snow upwards. Maybe not directly up, but close enough.

Re:I wish they'd do it here.

[AK Marc]

I've read that streetlights don't help with traffic crashes at all (crash rates in lit areas is roughly the same as unlit areas). But people cling to things emotionally, not logically. Streetlamps for pedestrian safety reduce safety, and speed bumps increase traffic crashes and reduce safety, but people will insist on both because their (broken) common sense says it should be best.

Re:I wish they'd do it here.

[cyberjock1980]

I admire your "plan ahead" approach, but in 2 small cities I've seen where they had some that were solar powered, they ALL were damaged by the accompanying natural disasters so they really didn't help even when disaster strikes. :(

They are just too fragile to hope to survive things like hurricanes, tornadoes, really bad thunderstorms, and earthquakes.

Re:I wish they'd do it here.

[Anonymous Coward]

How many street lights do you see on the 7th floor? Also many have heaters now to combat it (if needed).

Re:I wish they'd do it here.

[EvilSS]

I've seen that with regular street lights.

Re:I wish they'd do it here.

[AK Marc]

Only when the dumb midwesterners do it. In Alaska (known for snow) the shroud was re-designed slightly to improve airflow over the light-face to reduce sticking.

Re:I wish they'd do it here.

[Samantha Wright]

I'm not sure where you're getting the "expects redder colors" part from. The Purkinje effect simply describes the fact that we're more sensitive to blue light at lower intensities—we see it better. This is purely physical, and due to the assymmetry in the response curve of all of our photoreceptors. While most direct light sources activate the cone receptors, this bias is sufficient to make us think of our monochromatic rod cell night vision as slightly bluish, which is why nighttime scenes are depicted as being blue in art, even though you're literally only seeing something grey. Rod cells have such a wide response range in the blue portion of the spectrum (not shown on graph) that some people can see very violetish frequencies with them, causing eyestrain as we get indecisive about how to dilate the pupil.

Sodium lamps are extremely monochromatic; they only put out a very small range around 600 nm because of the chemical reaction that they operate on. Any white bulb either incandescent or LED, even ones with a bluish tint, will illuminate red signs much better than a traditional sodium-vapour lamp.

Re:I wish they'd do it here.

[amRadioHed]

Icicles can and do form on all kinds of overhead surfaces in the city. They do present a small risk, but that has nothing to do with the streetlights.

Re:I wish they'd do it here.

[Samantha Wright]

It's also believed to be why programmers become nocturnal; the pure white light of a computer monitor screws up part of the Circadian rhythm. It's quite possible that all of New York will become even more insomniac after this change. The blueness of the light is surprisingly important; pure blue light is around four times more potent than white light in treating seasonal affective disorder, a form of depression caused by lack of daylight.

Re: I wish they'd do it here.

[jd2112]

Too late. NYC has been called "The city that never sleeps" for decades.

Re:I wish they'd do it here.

[suutar]

In my experience streetlights aren't helpful for seeing traffic, but they are very helpful for seeing the path of the road. Once on a road trip I was out in the middle of Wyoming at night and it was disconcerting to be unsure of the edge of the road; I had to pay more attention to the section right in front of me and thereby pay less attention to what was coming up further ahead.

Re:I wish they'd do it here.

[WaffleMonster]

Anyone out there with more specific info and/or why one color is better than another for night vision preservation?

All light u can see destroys night vision. Rods (e.g. peripheral vision) are not as sensitive to red light and take much longer than cones (e.g. center, detail, reading vision) to regain their low light sensitivity. Cones are what allows you to see with any detail..your legally blind without them and so if they are not sensitive to a frequency that frequency is not doing you much good.

The takeway is intensity rather than color is king if you need to preserve your night vision. Color games might be worth playing on a moonless night on the high seas...but they are just as likely to screw you over if the environment is unsuitable (e.g. red lights while reading maps with red warning labels don't mix)

Re:I wish they'd do it here.

[Anonymous Coward]

You don't need filters or phosphors for street lamps. Just group red, green and blue leds tightly. The resulting mixed light will be white enough when it comes all the way down to the road. Street lamps can have more coarse mixing than a desk lamp, because they are much farther away from what they illuminate.

Re:I wish they'd do it here.

[kaatochacha]

Is there an unofficial law, similar to Godwins, that applies to people dragging partisan politics into completely unrelated discussions?

Re:I wish they'd do it here.

[plover]

The thing is night street lighting decisions have rarely been dominated by discussions of color. They always boil down to cost.

When mercury vapor lamps were invented, they were more energy efficient than incandescent and fluorescent, and they became the dominant street lighting technology despite their ugly blue color. As metal halide lighting improved efficiencies, (and color,) they began to displace mercury lamps. High pressure sodium came along, and was even more efficient. Cities installed them despite the distinctive golden glow that made them bad for color rendition. (It turned out to be good for avoiding light pollution.)

LED lighting is actually not as efficient as HPS. (68 lm/watt vs 100 lm/watt). Where LEDs offer a lot of benefit is in lamp life. A 145 watt LED streetlight costs $1100 for the fixture and electronics, costs $200 to install, produces 10,000 lumens, should last 50,000 hours, and the power cost is about $1100, for a total of $2400 for an 11 year service life. A 400 watt HPS or metal halide or sodium vapor bulb produces 55,000 lumens, but has to be replaced after 25,000 hours. So it costs $200 for the fixture, $30 for the bulb, $200 to install, $30 for a second bulb, $100 to change it, and $3000 for the electricity, for a total of $3560.

They could tint the color more to the red spectrum by using a different choice of phosphors on the LED chips, at a cost of reduced light output.

Most cities seem willing to settle for the lower light levels in exchange for the substantially reduced overall costs.

Re:I wish they'd do it here.

[Godai]

I think you're confusing street lights and traffic signals. Places in the snow belt have had issues with LED traffic signals getting blocked with snow, but I can't see the same thing happening with a downward facing street light.

Lord, I should hope not. If the snow's that high I think 'blocked street lights' is the least of your trouble!

Re:I wish they'd do it here.

[Mashiki]

I wonder how many smaller cities have already done this?

In Canada it's common, the problem that we find up here though is the LED street lights don't generate enough heat to keep the cover clear in the winter. Causing severely reduced visibility during heavy snowfalls. This of course is leading to heating units being installed to melt the snow off, so there's enough light to make it safe. And it pretty much negates any energy savings. My hometown messed around with a few test areas on major streets for two winters, and then switched back to pressurized bulbs. Especially after those stretches of road became known for accidents.

Re:I wish they'd do it here.

[Anonymous Coward]

Our local utility did it here, small town of about 4K population in WA about a year ago and IMHO the difference is positively dramatic. The light seems more directed and easier on the eyes with as much or more clarity. I highly recommend it.

Re:I wish they'd do it here.

[Anonymous Coward]

Chandler, where I live, we have all LED's in the neighborhoods. I like them, actually.

Re:I wish they'd do it here.

[Darinbob]

We have a few varieties of LED streetlights in the office, and they're extremely bright. I was surprised that a handful of LEDs could pump out that much light.

Re:I wish they'd do it here.

[Alef]

If you want daylight-like colour, there are other alternatives than LEDs. Ceramic metal halide lamps, for example, have excellent colour rendering and about the same efficiency and life expectancy as LEDs, at a significantly lower cost. The main drawback is that they take a while to fire up, but that isn't really a problem with street lights.

Re:I wish they'd do it here.

[Anonymous Coward]

I think Dumbledore will still find a way around that when he needs to protect young wizards.

Re:I wish they'd do it here.

[Darinbob]

And these LED lights are warm. Much of the heat is from the metal on the top of the fixture. Granted not as hot as incandescent but they'd definitely melt snow.

Re:I wish they'd do it here.

[Anonymous Coward]

Why the fuck are you comparing a 145W LED to a 400W HPS?
Do the same math with a 150W CMH or HPS:
$200 for the fixture, $30 for the bulb, $200 to install, $30 for a second bulb, $100 to change it, and $1150 for the electricity, for a total of $1710.
Whoops.

Re:I wish they'd do it here.

[smooth wombat]

Streetlamps for pedestrian safety reduce safety

Bullshit. Street lighting has been found to reduce pedestrian crashes by approximately 50%.

peed bumps increase traffic crashes and reduce safety

Double bullshit. Overall, the treated streets experienced a 39 percent decrease in crashes per year after speed bumps are installed. The 39 percent decrease on speed bump streets is a statistically significant difference (t = 2.8) from 1.39 to 0.85 crashes/year, meaning crashes most likely do decrease on speed bump streets due to bump installation. As well as this gem which asks a different question but which provides the same evidence against your "common sense".

Re:I wish they'd do it here.

[Anonymous Coward]

Ours just have an opening at the bottom of the sun shade cylinder. Prevents bird nests too. Easy peasy!

Re: I wish they'd do it here.

[adolf]

According to what I've read, NiFe batteries can last darn near forever (as far as batteries go), but they do require maintenance: The electrolyte needs occasional topping-off and replacement.

Furthermore, they're expensive.

According to these folks, the smallest 12V package is $1010.00, and consists of ten 6x3x15 cells of 15 pounds each. It provides 100 Ah of 12V power.

Which is roughly enough to run a single streetlight all night, assuming that the streetlight draws around 100W (which I think is a reasonable assumption). (100 Ah * 12V = 1,200 Watt-hours / 100 Watts == 12 hours runtime, ish.)

Meanwhile, a quick Google search shows that a 100Ah deep-cycle lead acid costs around $200.

Is NiFe a better value at around 5 times the initial cost, factoring maintenance requirements? It all depends...

Re: I wish they'd do it here.

[onepoint]

ROFL
thank you I need that laugh

Re:I wish they'd do it here.

[nospam007]

"We have a few varieties of LED streetlights in the office, and they're extremely bright. I was surprised that a handful of LEDs could pump out that much light."

Wow, an office so big, that it needs streetlights.

Re:I wish they'd do it here.

[nospam007]

"I don't think I've ever seen it snow upwards.

Never seen a good blizzard?"

Not to mention Australia.

Re: I wish they'd do it here.

[FishTankX]

Well, the topping off of electrolyte and potash is every 13 years. This might be able to be be automated. But given careful stewardship i wouldn't be surprised if a large lead acid battery using shallow depth of discharge could be cost effective in the short term. However, all things considered, when you're talking about the span of 50+ years I think the NiFE batteries would win out, if anything due to disposal costs, because lead acid batteries would need to be carefully recycled. With NiFE batteries there are no seriously toxic constituent chemicals, so you could probably just leave them buried there if they die, or service them and press them back into another 50+ years of service. But since there are some original Edison cells kicking around now from around 80 years ago, I'd imagine they'd outlast their maintainers.

However, in the long run it may just be more prudent to have one backup diesel generator running the street than a row of expensive batteries. If you've got 500 street lights to power, a 50kw natural gas generator could be deployed for about $10,000. Assuming your battery backup + solar cell system only costs $250 a light, the natural gas generator is still 1/12th the cost. It could run on a separate circuit, to avoid being hazardous to repair crews on the power grid, and run until morning where it could be shut off and attended to. Given that it would only need to run maybe once or twice a year, you'd have a looong service life, and it could be plumbed directly into municipal gas lines to eliminate the issue of stagnating fuel.

Re:I wish they'd do it here.

[nightsky30]

Yes, we've had LED traffic signals here for years, and I've only seen them obstructed by snow once. You need a wet, sticky snow and a swift drop in temperature for it to happen. IINM they put remote-controlled heaters in the newer ones.

And it seldom snows upwards. I don't think I've ever seen it snow upwards.

Check the image in this article from a few years ago...
http://www.huffingtonpost.com/2009/12/16/led-traffic-lights-that-c_n_393769.html

Re:I wish they'd do it here.

[onepoint]

when I worked downtown / wallstreet. the walk from the Path station ( old WTC ) to the exchange had
rain going up
snow going up
and skirts going up

always in the winter

it's common and you do see snow blowing in all directions, even on the 23rd floor
( rain when it goes up is really weird and you would swear you were in some sort of
movie slow motion set )

Re: I wish they'd do it here.

[adolf]

Disposal costs? Lead is very easily and profitably recycled. The last battery I sold was a decent-sized SLA; I got $9 and some change for it at a neighborhood scrap yard. It will be turned into new car batteries or somesuch. If you think lead is expensive to dispose of, you're doing it wrong.

The generator is a bad idea. Fuel gets old, fluids need changing (even if it is never actually used), engines need run periodically, they have their own batteries to maintain, and etc. Standby generators are not cheap propositions, if there is any expectation that it actually work when needed.

I think the whole thing is being vastly over-thought, anyway: So the power goes out, and it is dark. So what? Do the raccoons, bats, and neighborhood housecats need light? It's cheaper, safer (no candles need burning, no houses catch fire), and more useful to give everyone a cheap LED flashlight, if having light outside is so critically important to human existence.

And I'm not sure that it is.

There are well-populated neighborhoods that don't have any street lights all over the country, and mayhem and chaos don't seem to continuously erupt there. For that matter, the town of Port Clinton, Ohio, purposefully goes into blackout mode for a few weeks every year when the mayflies are at peak, and this darkness doesn't seem to make Port Clinton any crazier than it normally is.....

(Amusingly, before they started doing this, they ran snowplows to literally debug the streets every morning: The little fuckers and their gooey entrails would pile up under each and every visible light.)

Re:I wish they'd do it here.

[Anonymous Coward]

I hope they're including the carbon emissions of manufacturing the new bulb when they say they're going to cut it 30%!

Re:I wish they'd do it here.

[Anonymous Coward]

Sleet can be a bit sticky sometimes, and as it drips down it freezes hard. With incandescent lights, they would stay warm enough where that would just melt and keep dripping off. I figure the LED lamps would become lightsicles with much reduced effectiveness. Under the conditions where freezing rain takes out trees, light fixtures that allow such buildup might be subject to the same kind of problems.

To be effective in locations with cold weather climates, the light fixtures (just like traffic lights) should have some of resistive heating element that turns on when it gets below freezing. Even though that would use a more electricity and reduce the total efficiency, it's still probably uses a lot less electricity than a high wattage incandescent fixture.

Re:I wish they'd do it here.

[Marillion]

In 2005, the Commonwealth of Kentucky replaced all 77,000 traffic lights in the entire state with LEDs over the course of about a month. Citation

Re:I wish they'd do it here.

[Anonymous Coward]

But when you have a little snow holder (as on traffic lights) that 20-30% efficiency doesn't generate enough heat to melt snow. Older incandescents would certainly melt the snow as they double as heaters.

Re:I wish they'd do it here.

[wonkey_monkey]

Stupid democrat. Go back to Russia!

Re:I wish they'd do it here.

[dbIII]

They are just too fragile to hope to survive things like hurricanes, tornadoes, really bad thunderstorms, and earthquakes.

Yes, bricks are too fragile in that situation. What you meant something else?
Such a comment as the one I've quoted is so easily applied to anything that it is entirely meaningless, so that makes me assume that it was only made due to blind hatred of the solution proposed.
Get over it, it's not 1960 any more and solar is practical in many situations.

Re:I wish they'd do it here.

[illtud]

This town in West Wales (Aberystwyth) has already done this on main roads - not just the lamps but the posts and housings as well. The're a great improvement - I'm about 100 yds from the trunk (main) road here as I type and I have to look closely to see if they're still on. The new lights are very directional (no light pollution) and they have a much better spectrum if you're walking along the road and the illumination seems much brighter. They replaced low-pressure (orange) sodium.

I'm sure that replacing the posts (6m ish tall) cost quite a bit, and I don't know what the efficiency saving is, but the payback must pretty short-term as the Welsh government doesn't have money to burn. It's been a great improvement here.

Re:I wish they'd do it here.

[Anonymous Coward]

Depends on the geography. Solar is fine for places that have enough sunshine year round to benefit (anything south of 45 degrees), lack wind and hail storms (which is pretty much nowhere, again depends on the degree of damage)

Places like Vancouver BC, Seattle, Portland and San Francisco are more likely to be damaged to destroyed by mega earthquakes that might only happen every 300 years, so installing solar primaries is viable. However they occasionally get a wind or hail storm that knocks tress over, in which the Solar powered LED lights can still work, albeit at reduced capacity if they're damaged or covered.

The problem for LED lights themselves is that they don't produce enough ambient heat, so cold places like Toronto and Montreal that get large snowfalls, the LED lights get blocked by snow collecting on them. A partial solution for this to design the LED bulbs heatsinking to be integrated into a peltier cooler/heater that can be flipped on to control the temperature and melt snow collecting in front of the bulb. The other partial solution is to not have the "shroud"'s on the bulbs so that there's no place for snow to collect, but this again requires a redesign of the entire light.

Places like New York City, likely get wind and snow, but nothing quite as destructive as tornados, at least not at the current rate of global warming/climate change.
They could integrate solar panels onto the structures the lights are attached to and recover some energy during the day, however I think it would benefit NYC more to just require buildings to have some percentage of solar panel cladding on the south-facing exteriors.

   

Re:I wish they'd do it here.

[snsh]

Those yellow low-pressure-sodium lamps are the most efficient source around. You see them around Palomar because they create far less less pollution and sky-glow than other sources. Astronomers love them, and they create less glare too.

Re:I wish they'd do it here.

[AK Marc]

The result is, if you can't see the road, you slow down. If you can see the road, you speed up. So the less light, the slower the drivers. If there are any unlit areas in an otherwise well lit area, people don't slow down for just that one spot (they can see the road beyond just fine), so a pedestrian or pet in the road at that point is more likely to get hit by a faster car.

So yes, in violation of common sense, fewer lights make for a safer road.

Re:I wish they'd do it here.

[AK Marc]

Street lighting has been found to reduce pedestrian crashes by approximately 50%.

Yes, in studies that compare pedestrians in unlit crossings to lit crossings, not by comparing pedestrians hit anywhere in an unlit city to pedestrians hit anywhere in a lit city. Your cite also says:
"The major criticisms of street lighting are that it can actually cause accidents if misused, and cause light pollution."

The issue is that most implementations are borderline "misuse" and thus can cause, rather than help crashes. Making sure a small test area is lit better than a wider rollout would get, and studying an unrealistic scenario often does get the expected results. Doesn't make it true.

Overall, the treated streets experienced a 39 percent decrease in crashes per year after speed bumps are installed. The 39 percent decrease on speed bump streets is a statistically significant difference (t = 2.8) from 1.39 to 0.85 crashes/year, meaning crashes most likely do decrease on speed bump streets due to bump installation.

http://www.motorists.org/traffic-calming/ Try reading something not written by people with a vested interest in increasing regulations.

Re:I wish they'd do it here.

[mirix]

Your image is for low pressure sodium, which is indeed monochromatic and very yellow. I haven't seen any in use in years, except industrial things where you need the absolute most light at least cost (LPS is even more efficient than LEDs), at things like big compounds, jails, etc. I remember some cities using them as street lights when I was younger, though.

Settlements near observatories tend to use them, as the single color makes the light easy to filter - and the lack of blue makes it scatter much less, too.

Most street lights are high pressure sodium - much whiter light, but still somewhat to the yellower end of things, as it has little or no blue emission. Very different! Less efficient too, for what it's worth.

Re:I wish they'd do it here.

[Samantha Wright]

Ah, okay. Where I grew up there were a lot of low-pressure sodium lights, actually, until a thorough re-sidewalking in 2009 or so. For HPS I would imagine a red stop sign would indeed be still somewhat dark, given this other spectrum. (Also, off-topic, your sig is the best.)

Re:I wish they'd do it here.

[Anonymous Coward]

No you've got that backwards. Green and red lights are better for night vision than any other colors. Blue is the worst as it convinces our circadian rhythms that it's still daytime.

Re:I wish they'd do it here.

[Anonymous Coward]

The change in crash rate was next calculated for the treated streets. Based on the data provided, it was found that there is a five percent reduction in crashes per ADT after speed bumps were installed. This change is not statistically significant (t= 0.22). This is different than the 39 percent reduction in crash frequency shown in Figure 27. The only difference between the crash frequency and crash rate is that the crash rate takes into account that volumes on the treated streets decrease after speed bumps are installed. Thus, of the 39 percent reduction in crash frequency on treated streets, the reduction in traffic volumes is the main factor in the reduction of crashes on treated streets. Fewer cars mean less potential for crashes.

So people tend to avoid roads with speed bumps - I'm not certain that was the effect you were trying to imply.

Re:I wish they'd do it here.

[Maury Markowitz]

> Yes, in studies that compare pedestrians in unlit crossings to lit crossings

So, "good" studies then.

> not by comparing pedestrians hit anywhere in an unlit city to pedestrians hit anywhere in a lit city

Given the paucity of "unlit cities" that might be used apples-to-apples, I'll stick with the studies I have to the ones I don't.

> it can actually cause accidents if misused

Note the term "misused" and the lack of the term "pedestrians".

Also note that the only referenced statement in the entire section has to do with stray voltage.

Also note that if you look up any of the unreferenced claims made in this section, the only hits you'll find are people quoting this article.

In fact, I'm going to mark it up now and take it to the talk page.

> Try reading something not written by people with a vested interest in increasing regulations

Instead, read something written by people who have a vested interest in the opposite, and brag about it on their web page. Yeah, I'll get right on that.

Re: I wish they'd do it here.

[Maury Markowitz]

> So the power goes out, and it is dark. So what

So the major injuries due to collisions between cars and everything else goes up by about 100%.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2564438/

I used to live in downtown Toronto and then moved to the burbs. I can say with no hesitation that I never had problems seeing people crossing roads downtown. It is downright scary out here. I've had numerous occasions where someone suddenly appears in my headlights less than a second away. 3M reflective tape works wonders, but perhaps 5% of people have that on them at night.

Solar powered lamps at intersections and crosswalks are an *extremely* good idea. The only question is whether or not to have the panel, or just keep the battery fresh using the grid. Today today a full-sized ~255W panel is around $200, a MPPT charge controller with lighting control and remove sensing about $175, and NiFe and LiFe batteries with 20 year life cycles about $1200. The total is far less than the price of getting an electrician to wire into the grid, which costs a minimum of $2000 *not* including parts and labor.

Re:I wish they'd do it here.

[Maury Markowitz]

> They are just too fragile to hope to survive things like hurricanes, tornadoes, really bad thunderstorms, and earthquakes.

So are power plants and distribution systems. The first thing that goes in these situations is the power.

Re:I wish they'd do it here.

[Maury Markowitz]

> In Canada it's common, the problem that we find up here though is the LED street lights don't generate enough heat to keep the cover clear in the winter

Ummm, the cover is on the bottom of the lamp. Not a lot of snow there. And yes, I'm Canadian.

Street *signals* might have problems with warm sticky snow, but studies demonstrated this was not a problem in actual use. The only example was runway lighting, but this was addressed by designing the system to take heat from the heat sink on the electronics and routing it to the luminare. In fact, the new PAR20's I purchased last week did something similar to re-route hot air back out of the socket. All passive.

> And it pretty much negates any energy savings

Right, because heating the lamps for a few hours 20 days a year would offset the energy use of conventional lamps, which heat them the same amount for 365 days a year.

Do you really think about your statements before posting them?

Re:I wish they'd do it here.

[Maury Markowitz]

> which could be a safety hazard.

Could?

http://www.lumec.com/newsletter/architect_03-10/Snow_storms_and_outdoor_LED_lighting.html
http://ledsmagazine.com/news/7/1/4
http://www.allledlighting.com/author.asp?section_id=3040&doc_id=559911
http://boingboing.net/2009/12/17/led-traffic-lights-d.html
http://www.cbc.ca/news/canada/prince-edward-island/led-traffic-lights-trouble-in-snow-1.1311220

Long and short, the number of such incidents is tiny, and the cost of fixing it so small, that few municipalities (if any) bothered installing active fixes, and even sending out crews to brush the snow away was still far less expensive than fixing incandescent versions that burn out all the time.

It always astonishes me how people will take the smallest possible issue, and then think that its more important than all the advantages. I'm glad no one listens to nay sayers, otherwise we wouldn't have Google or the Internet to read all the whining.

Re:I wish they'd do it here.

[vtcodger]

Not a bad idea, but the schedule may be kind of aggressive. It doesn't seem to allow a lot of time for dealing with problems.

I'm aware of at least one outdoor LED roll-out that hasn't been problem free. It's at the recently rebuilt (at a cost of $76,000,000) Crown Point Bridge over a narrow spot on Lake Champlain. It's not that big a deal since drivers crossing the bridge at night have their headlights on anyway. But similar problems in NYC would presumably earn some bad publicity and increase costs beyond what is expected.. Here are some links http://www.northcountrypublicradio.org/news/story/21643/20130320/why-don-apos-t-all-the-lake-champlain-bridge-lights-work http://www.rutlandherald.com/article/20130628/THISJUSTIN/702149975

The problem apparently isn't the LEDs themselves. It's the circuits powering them.

Re:I wish they'd do it here.

[Hognoxious]

our visual cortex expects redder colors in darkness

I don't know if the visual cortex expects anything, but the eye is more sensitive to blue at low levels of illumination. "Starlight" happens to have a blue bias. Some say that might not be a coincidence.

Are you thinking of the red interior lights in war films, before the commandos/seals etc go outside? That's to avoid buggering up your night vision - the rods & the blue cones don't react to it and get bleached out.

Re: I wish they'd do it here.

[adolf]

$2000, not including parts and labor?

What is the $2000 for, if not parts and labor? Hookers and blow?

Re:I wish they'd do it here.

[SageMusings]

And they are already beginning to burn out. There are very many traffic lights in Louisville with as much as 20 - 30 % dead cells in the matrix the driver sees.It looks like a diminishing patchwork of little squares. I was quite surprised at their true lifespan. Perhaps that is still far superior to the old lights but I don't have any data. It is, however, easy to see how the LED lights are aging.

Re:I wish they'd do it here.

[sayno2quat]

I don't think I've ever seen it snow upwards.

It comes up, Charlie Brown, snow comes up! ~ Lucy

Re:I wish they'd do it here.

[Anonymous Coward]

The white ones are now more efficient than any other white light source on the market and the price has dropped to half of what it was 3 years ago. There's never been a better time to switch if you have significant lighting needs. High pressure sodium (what they are replacing) was the most efficient white source available until last year. The only light more efficient for street lighting is low pressure sodium which was abandoned for being a narrowband yellow and unpleasantly difficult to see by.

Costs

[sfm]

Yes, there is a savings, but how much is it going to cost NY taxpayers up front ?
Would a better strategy be to replace the sodium lights with LED style lights, as they wear out?

Re:Costs

[mythosaz]

They're doing that *too* of course, since while they're rolling them out, they'll still need to replace broken ones in parts of town that they're not in.

...but the question is simply a matter of if it's more efficient to replace an existing, working, lightbulb with an LED.

I know it was for me, but my replacement costs were low. [I didn't have to pay the laborer...]

Re:Costs

[Anonymous Coward]

Exactly, but we know there are no smart people in politics, only guys with their hands in everyone else's pockets.

Re:Costs

[Anonymous Coward]

Not if the power draw between the two is high enough, electricity costs money

Re:Costs

[Anonymous Coward]

Nope, its actually cheaper to replace them in large blocks than to replace them one at a time. Old tech told me once that when they maintained the long tubes at the factory high up. Once a few went it was only a matter of about a year or two before the rest of the them did and it was more disruptive, time consuming, and costly to replace them one at a time than to do it all at once. So, I've followed this process for most of my larger lighting projects. If you are going to replace one brake light replace them all. If you are going to replace one headlight replace them both.

Also, the nice thing about LED lighting is that the way it fails is it just doesn't produce as much light as it once did. So in 20 years if they want to put off the costs for another couple of years, it's entirely possible to do so.

Re:Costs

[Anonymous Coward]

Nope, its actually cheaper to replace them in large blocks than to replace them one at a time

I agree. I'm not going to try to run the numbers, but you have one unionized cherry-picker crew retrofitting a block in one day vs. many of the same crews going out on job, after, job, after job on the same block. There's a fixed cost to rolling out on site, so replacing them all at once definitely seems like the cheaper way. This is purely intuitive though.... it could still be wrong.

Re:Costs

[M. Baranczak]

Their replacement costs are probably much higher than the cost of the bulbs. So it makes sense to pay extra for bulbs that last longer, even if you ignore the efficiency gains.

First, they have to send a guy out there. And for most of the lights, this guy will need a truck with a cherry-picker. And if it's a single-lane street, then the truck will be blocking it during this operation, so you need to redirect traffic. And probably a bunch of other weird shit that only applies to NYC.

Re:Costs

[SuperBanana]

Yes, there is a savings, but how much is it going to cost NY taxpayers up front ? Would a better strategy be to replace the sodium lights with LED style lights, as they wear out?

Upfront costs don't matter much; the cost is in the labor getting to these lights. Every time they replace one, the future labor cost drops.

Re:Costs

[Darinbob]

There are also plans in many place to get these lights onto a network so that they can be monitored as well. Often the city doesn't know that the bulbs have burned out until people start calling, whereas they can be more proactive by measuring current use and detect failing units sooner. Also there is more control over when to turn them on and off, schedule them for dusk and dawn, etc. A lot of places don't even have photocells attached to the lights to know when it's dark and with a fixed timer schedule that doesn't vary.

Re:Costs

[Anonymous Coward]

Except for the fact that the working life of of LED bulbs seems to be a fraction of the life the individual LEDs themselves would have if you fed them nice, clean, properly-regulated DC with proper heat management.

I've bought 6 LED bulbs in approximately four years.

One literally went up in smoke on day 3.

One lasted about 4-6 months.

One lasted about 7 months

One lasted about a year.

The other 3 still work, but I'll be amazed if they're still burning 5 years from now, let alone 16. Lifespan-estimates for LED bulbs are a complete fraud. They either quote the ideal lifespans of the LED elements under test conditions, or just lie. The problem isn't the LED elements, it's the crap components soldered to the same circuit board (or lack of components to buffer them from things like voltage surges).

Re:Costs

[ebvwfbw]

If you are going to replace one brake light replace them all. If you are going to replace one headlight replace them both.

Also, the nice thing about LED lighting is that the way it fails is it just doesn't produce as much light as it once did. So in 20 years if they want to put off the costs for another couple of years, it's entirely possible to do so.

That's not right. Replace just that bad headlight unless the other one is so old it has had it or money is burning a hole in your pocket. I've never had the other headlamp go bad shortly after the first one. Maybe years later and I think that happened once in 40+ years. Now if the package comes in a two pack, yea I replace the other one as well because then it's in the garage and you'll never remember what it went to.

This is also true with CFL bulbs. I've kept track with my rentals. Some blow within 6 months, one has been burning for 10 years so far. These are outdoor lights that burn all night long. They last longer than incandescent bulbs. So far I haven't installed a LED light for an outdoor application.

So don't go and replace them all unless you want to. You're not saving yourself anything.

Re:Costs

[Anonymous Coward]

I like my headlights on alternating outages. Yes you need to run to the auto-parts store more often, but there's less chance of both lights going out at the same time.

When do the unions go on strike ...

[drnb]

The LEDs have a lifespan of 20 years, more than three times that of the current incandescent bulbs, and Bloomberg says it will save $6 million in energy and $8 million in maintenance every year.

When does the electrical workers union go on strike, the bulb changers belong to this one, followed by various other city worker union going on strike in support?

Re:When do the unions go on strike ...

[mythosaz]

Well, not *just* yet, anyway.

I'm sure there's plenty of kickbacks in the roll-out phase.

So... ..3-5 years?

Re:When do the unions go on strike ...

[Anonymous Coward]

Probably about as quickly as the plumbers' unions in Philadelphia went on strike when they learned the new Comcast building would have all waterless urinals in it. They sued, won, and forced Comcast to pay them to install miles of water that wasn't hooked up to anything. It cost Comcast subscribers millions of dollars.

Re:When do the unions go on strike ...

[mcgrew]

The IBEW won't strike over this any more than they went on strike to prevent the use of bucket trucks or remotely-read electric meters; my dad was a lineman, and for the first half of his career they had to climb the poles. Changing the bulbs in streetlights is just a tiny part of what they do, more often they're changing transformers squirrels blow up when they crawl inside to get warm.

Re:When do the unions go on strike ...

[necro81]

When does the electrical workers union go on strike, the bulb changers belong to this one

We'll just ship them off to a different planet, along with all the telephone sanitizers.

[obscure reference?]

Re:When do the unions go on strike ...

[Anonymous Coward]

> It earned Comcast executives millions of dollars.

There. Fixed.

Re:When do the unions go on strike ...

[Anonymous Coward]

I wish.

Re:When do the unions go on strike ...

[Anonymous Coward]

Probably about as quickly as the plumbers' unions in Philadelphia went on strike when they learned the new Comcast building would have all waterless urinals in it. They sued, won, and forced Comcast to pay them to install miles of water that wasn't hooked up to anything. It cost Comcast subscribers millions of dollars.

I'm no eco-zealot, but I try to moderate my footprint on our planet. That being said, fuck waterless urinals and their stench of rancid piss. If I wanted to smell that I'd have pissed on that homeless guy in the alley.

Re:When do the unions go on strike ...

[DexterIsADog]

Comcast is just one highly visible example. However, I didn't find any reports that the unions sued anyone. The city board of licensing and inspection did vote on whether to approve the building, and the union probably complained to them to request a ruling requiring the pipes. The building developer's spokesperson did say, "It was always our intention to run the additional pipes." Do you have any link that mentions lawsuits?

The urinals in my building are waterless, but each one has a capped water supply pipe sticking out of the wall above them. NJ Strong, indeed.

Re:When do the unions go on strike ...

[Anonymous Coward]

When the urinals and all the drain piping get crusted up with pee cheese, someones going to have to put real urinals in. Just about everyone who install those damned things get sick of the bathrooms smelling like an outhouse.

Re:When do the unions go on strike ...

[Anonymous Coward]

Whether or not the plumbers union is right or wrong, a tiny little bit of critical thinking would indicate the whole thing probably did not cost millions of dollars extra for Comcast.

For one: each urinal will have a drain; no doubt much the same as urinals which do use water. Each restroom will already have faucets, toilets which use water. The extra cost in materials and labor to run one capped cold water stub per urinal up into the wall, done in the least efficient way possible couldn't cost more than a thousand dollars extra per restroom, even at union wages.

On the other hand, if the offices get tired of the inevitable piss stink, and decide to install regular urinals, the one sure thing is the plumbers union just saved them hundreds of thousands in remodeling costs.

Re:When do the unions go on strike ...

[Anonymous Coward]

Piping that would allow water to flow to the urinals in case they needed to be converted was installed in the Comcast Center, which Liberty Property Trust says was always part of the building's plan.

https://en.wikipedia.org/wiki/Comcast_Center_(Philadelphia)

incandescent != sodium

[Anonymous Coward]

"high-pressure sodium street lights" are not incandescent bulbs.

Re:incandescent != sodium

[EvilSS]

Great, problem solved! I'll call the mayor and let him know he can cancel the project...

Re:incandescent != sodium

[MightyYar]

Came here to say this. Sodium lights are already pretty darned efficient, if a bit ugly. Sodium lights get 140 lumens per watt, so I'm not sure where their savings are coming from - perhaps the improved quality of light lets them decrease the lumens.

Re:incandescent != sodium

[Anonymous Coward]

Sodium lamps START their lifetime at around 140 lumens per watt. They get progressively inefficient as they age. While they will still put out a similar illumination over their lifetime, their power usage progressively climbs.

Re:incandescent != sodium

[SJHillman]

I did a quick Google to satisfy my curiosity and found a few things:
- While high pressure sodium gives off more lumens per watt, LED has better effective illumination (in part due to how our eyes can detect different wavelengths)
- LEDs are more directional, eliminating up to 40% of light loss due to reflectors
- In the end, an LED might only need to give off 20 or 30% as much light to still illuminate the same area effectively

Source: http://www.al-e.com/led-vs-sodium-lamps

Re:incandescent != sodium

[kyrsjo]

Unfortunately, so does LED. They tested a similar rollout in Oslo recently, but had to stop when they discovered that the LEDs aged far more rapidly than the old sodium fixtures.

Source:
http://www.abcnyheter.no/nyheter/2013/06/17/her-er-grunnen-til-oslo-satte-full-stopp-pa-led-utbyttingen

Re:incandescent != sodium

[umafuckit]

In the end, an LED might only need to give off 20 or 30% as much light to still illuminate the same area effectively

Source: http://www.al-e.com/led-vs-sodium-lamps

Yeah, but it looks like what they're actually doing is over-lighting. So there's way more light than before. It's a disaster for amateur astronomy, nocturnal species, and even diurnal species which have their circadian rhythym disturbed.

Re:incandescent != sodium

[Anonymous Coward]

According to... wait for it... an LED lighting company.
    Now there's an objective, unbiased source of I've never seen one!

Re:incandescent != sodium

[MightyYar]

It's a disaster for amateur astronomy,

You mean the one star that I can sort-of see in NYC will disappear? :)

I think that damage is already done. My daughter didn't really know what a star was until we brought her to the beach. After that, I felt pretty sorry that I had been singing "Twinkle Twinkle" all this time without actually telling her what the heck a star was...

Re:incandescent != sodium

[smellsofbikes]

Another advantage, if purchasers care to implement it, is that you can have somewhat intelligent LED lights that dim down to 30% when there's no traffic around, so it's still light, but much lower power, then run back up when traffic is a block away. It doesn't add much to the system cost to add motion detection and communication with nearby lights, particularly since some industrial/commercial LED lights are adding selftest health/failure reporting already.

Re:incandescent != sodium

[umafuckit]

It's a disaster for amateur astronomy,

You mean the one star that I can sort-of see in NYC will disappear? :)

No. I mean that suburban areas which now have a limiting magnitude of 5 or so (okish for astronomy but not great) will turn into NYC. Also, the skyglow from these areas will creep into currently pretty good skies. e.g. The skyglow from the NYC area is currently just visible on the SE horizon from the heart of the Catskills, but we still have some dramatic skies there (Milky Way almost to the horizon). With these white lights appearing and spreading, that is going to change.

Re:incandescent != sodium

[geoskd]

I did a quick Google to satisfy my curiosity and found a few things: - While high pressure sodium gives off more lumens per watt, LED has better effective illumination (in part due to how our eyes can detect different wavelengths) - LEDs are more directional, eliminating up to 40% of light loss due to reflectors - In the end, an LED might only need to give off 20 or 30% as much light to still illuminate the same area effectively

The part you were missing is the ballast. an old fashioned (non electronic) ballast uses just as much power as the lamp and produces no light at all, cutting the actual light produced per watt in half. Modern electronic ballasts still consume around 1/3 of the power the lamp does, so you're still cutting down your efficiency by 25%.

Re:incandescent != sodium

[mynamestolen]

yep. lots of theory about leds but many are a disaster when it comes to quality control (including some expensive ones I put in our house). Let's hope they can do better than the pro-business New Zealand government leaving it all to the market.
http://www.nzherald.co.nz/nz/news/article.cfm?c_id=1&objectid=11131680

Re:incandescent != sodium

[ed1park]

But will they attract more insects at night?

Re:incandescent != sodium

[Osgeld]

the problem with replacement "bulbs" in LED is they have to take a fairly expensive complicated technology and compete with non complicated established technology. Even the expensive bulbs use the absolute cheapest methods and materials they can get their hands on.

Having cheap LED's with the only thing between them and the electrical grid being a 2 watt resistor and a slug of what you hope is zinc is going to expose them to a freaking ton of abuse, and thus drastically shorter time periods. Doing it right would increase the cost of the bulb so much that every yutz in the universe will avoid the product and buy the first 9.99$ LED bulb shipped directly out of china with a phoney UL listing.

Re:incandescent != sodium

[evilviper]

Sodium lights get 140 lumens per watt

Strange that you went out of your way to say this, yet failing to note that there's a huge difference between low-pressure sodium and high-pressure sodium, and also not mentioning the efficiency of LEDs.

In-the-lab, LEDs are the most efficient form of electric lighting, demonstrated at 254 lumens/watt last year... better than HPS, better than LPS, and better than any others. The efficiency you'll get with inexpensive retail bulbs is significantly lower, though bulbs based on moderately expensive emitters (like a Cree XM-L2 U3) can exceed 150lm/watt.

http://www.cree.com/news-and-events/cree-news/press-releases/2012/april/120412-254-lumen-per-watt

http://www.cree.com/led-components-and-modules/products/xlamp/discrete-directional/~/media/Files/Cree/LED%20Components%20and%20Modules/XLamp/Data%20and%20Binning/XLampXML2.pdf

Re:incandescent != sodium

[evilviper]

they discovered that the LEDs aged far more rapidly than the old sodium fixtures.

I'd consider time-travel to be a bonus, rather than a drawback.

Re:incandescent != sodium

[MightyYar]

I'm not exactly a light expert :)

Low pressure sodium is indeed even more efficient - can we just agree that, as far as lumens per watt - sodium lamps are already pretty darned efficient? They may or may not manage to find LEDs with more lumen efficiency, but the color will almost certainly be better. I can see that (and their directionality) enabling them to use fewer lumens overall.

Re:incandescent != sodium

[evilviper]

No worries... I'm sure dozens of other folks said the same thing, some with more info, and many providing less. Yours just happen to be +1'd by some random moderator, and get more exposure as a result.

BTW, I'm no lighting expert, either. I've had a modest interest in them for quite a while, but recently I've been going through and upgrading the flashlights and indoor lighting for myself and my family, so I've read-up a bit on LEDs lately. I was tracking down replacements for an outdoor HPS light just the other day.

Re:incandescent != sodium

[Anonymous Coward]

I did a quick Google to satisfy my curiosity and found a few things:
- While high pressure sodium gives off more lumens per watt, LED has better effective illumination (in part due to how our eyes can detect different wavelengths)
- LEDs are more directional, eliminating up to 40% of light loss due to reflectors
- In the end, an LED might only need to give off 20 or 30% as much light to still illuminate the same area effectively

Source: http://www.al-e.com/led-vs-sodium-lamps

Since it's an LED company making these claims, it's basically an admission that sodium is more efficient, so they have an uphill battle selling LED's.

I don't really care about efficiency, so I don't think the rollout is a mistake, but it's not like we have any influence on whether it'll happen or not so the thing worth discussing is the lessons learnable from this engineering decision. "LED's are the most efficient kind of lighting" is FALSE. They are less efficient than sodium.

One reason I don't care about efficiency is that they draw energy only at night, so they contribute to base load which is basically free. It's less than 1/8 the cost of peak load. Right now, in NY, that means nuclear and coal. If, in two decades, we have some super-efficient renewable energy, it'll probably be wind or hydro, which cost based only on peak capacity and have no marginal cost for load.

Another reason is that I'm often out at night, so good lighting improves quality of life. I really appreciate the natural gas lighting in Frankfurt and Prague, for example: it's warm, but still incadescent white. It's not like the silly gas lamps we have in lower manhattan with a smokey orange open flame. They use a wick, like a camping lantern. However I worry I won't actually get better quality light from this retrofit. They'll use the ugly yellow+purple so-called-white LED's which will be worse than sodium. The focus of the article on "efficiency" of the light suggests a disregard for the right kind of light to use at night to make the city actually nice.

High-pressure sodium isn't "incandescent"

[Anonymous Coward]

It's a luminous vapor, rather than light from a heated solid filament. It's already much more efficient than incandescent bulbs to produce the same amount of light.

Re:High-pressure sodium isn't "incandescent"

[Anonymous Coward]

Canadian figures to back up my claim of efficiency

Re:High-pressure sodium isn't "incandescent"

[ebno-10db]

Why would I trust anything Canadians say? Underneath those obsequious manners and maple syrup, they're planning to invade America. Never let it be said the British Empire gives up easily.

Re:High-pressure sodium isn't "incandescent"

[thule]

Isn't it LOW pressure sodium? The low pressure ones have that orange look. They are extremely power efficient, but they give off a narrow spectrum of light. I have heard cops and paramedics hate them because it is hard to tell what is blood, oil, or water spilled on a street.

Re:High-pressure sodium isn't "incandescent"

[Anonymous Coward]

I hate them because when I park my car at night in a parking lot lit with them, the fact I recall that the rental car was "blue" really doesn't help me at all in finding it.

Re:High-pressure sodium isn't "incandescent"

[rubycodez]

you forget the beer, that keeps them docile. and schnockered.

Re:High-pressure sodium isn't "incandescent"

[evilviper]

Why would I trust anything Canadians say? Underneath those obsequious manners and maple syrup, they're planning to invade America.

Not to worry... US Beer makers have been anticipating the problem for decades, and tweaked the formula to be intolerable to Canadians. That should be enough to drive them back North, if they don't get distracted by the first hockey rink they come across...

Bout time

[Anonymous Coward]

I was wondering at what point Cities would make the full transition to LED lighting. Though, I expected them to switch over to color changing rather than single color lighting. Just as reliable, but gives better ability for safety reasons. Say if there was a gunshot in an area the colors could change to highlight the area for the helicopter pilot to immediately get drawn into or police officers on scene. Colors can also be used to increase the plant growth or blooming in specific areas like parks.

Re:Bout time

[Anonymous Coward]

It's a bold prediction, but in a few years I predict police (and even helicopter pilots) will have some sort of system using satellites to help them locate a specific location by its coordinates.

Admittedly, it's going to cost a lot to develop such technology but I'm going to make a proposal to develop a prototype system leveraging existing government infrastructure where possible. I think I can do it for under $10B which makes it a real bargain for taxpayers.

Re:Bout time

[Anonymous Coward]

It might be more useful to use different colors to communicate to "civilians".

For example, flash red on the path emergency vehicles are taking so drivers can pull right and stop (as they should) before they can figure out where the siren (if they can hear it over their music or cell phone) is coming from and/or see the vehicles' flashing red lights.

Of course, all such systems will only help at night if their cost is to be practical.

Current lights are NOT incandescent

[EmagGeek]

Metal-halide vapor lights are not incandescent.

Re:Current lights are NOT incandescent

[Anonymous Coward]

I bet they are and you would be too if you were about to be replaced!

Re:Current lights are NOT incandescent

[ebno-10db]

You're right, but they're replacing high pressure sodium vapor lamps (the ones with the fugly pinkish orange color). Metal-halide are bright white, take a while to warm up, and are usually used in places like school gyms. Occasionally I've seen them used in parking lots.

20 year lifespan

[Bayoudegradeable]

Let me first say that I live in New Orleans, so go ahead make all your inept government remarks now. That said, we did begin making changes in our traffic signals to LED lights and the big claim of "20 year lifespan" was made. Less than 5 years later I see many of the LED bulbs (really, clusters of bulbs, like a Lite Brite set) are now replaced with the traditional traffic signal bulbs. Not only did the LEDs not last very long, they aren't being replaced with LEDs but with the old style bulbs. Hope NYC gets LEDs from a better vendor than we did.

Re:20 year lifespan

[magarity]

Because traffic lights go on and off and on and off all day long. They're a terrible idea for LED because the ballasts wear out doing that. The actual LED built into the thing is still perfectly OK actually. Your local government (and many others) have been scammed by clueless but well meaning greenies on the city council over the traffic light LED thing. Streetlights that come on and stay on all night long are a much better use.

Re:20 year lifespan

[Anonymous Coward]

You must have bought the chinese versions. They've been pushing a cheaper version that's bout 30-50 percent less for a while now. The problem is that the life is literally 5-8 years instead of 15-20.

Re:20 year lifespan

[Megane]

The problem with LEDs in traffic lights, in my experience just driving around, is that the main problem is the light module itself somehow fails such that some of the LEDs can't get electricity. This results in part of the LEDs not lighting. New Orleans is right on the ocean, so the salt water in the air is more likely to cause corrosion problems. The equivalent would be blaming incandescent lights because the bases of the lamps fell apart after wind shook the signals around too much.

And then there is the problem up north, where incandescent traffic light lamps would keep the winter snow melted. When they were changed to LEDs, the lights started to freeze over from the lack of heat.

LEDs aren't perfect for traffic lights, but at least they're actually monochromatic.

Re:20 year lifespan

[Bratch]

Traffic signals are a different application. A few yeard ago I heard of a city somewhere up north that replaced all their incandescent traffic signals with LEDs. When winter came, the new signals, using 90% less power, didn't emit enough heat to melt the snow that accumulated in the signal housing. The snow can build up enough to completely block the light, resulting in confusion and accidents all over the city. They had to either go back to incandescents, install a heater element, or modify the housings. This shouldn't affect the street lighting.

All of our MTS busses have converted to LED headlights, which isn't hard to notice because they are exessively bright, even during the day.

Re:20 year lifespan

[khallow]

They've been pushing a cheaper version that's bout 30-50 percent less for a while now. The problem is that the life is literally 5-8 years instead of 15-20.

You should mean roughly 50-70 percent less. Because that's what you're claiming with the lifespan. That's a big difference.

Re:20 year lifespan

[ebno-10db]

Nobody said they worked underwater.

Re:20 year lifespan

[Kardos]

If that's true, then it's a terrible design. Only one of the lights is on at a time, right? So the ballast can run continuously, switching between red/green/yellow as appropriate.

Re:20 year lifespan

[Anonymous Coward]

A ballast is only needed to limit current for arc-based light sources. So fluorescent lights need them and mercury- or sodium-based street lights need them, but LEDs definitely do not. An LED lamp just needs a power supply and that's about it.

dom

Re:20 year lifespan

[foobar bazbot]

That's terrible for LEDs with circuitry designed for continuous operation. It shouldn't be an issue in traffic lights, because the LED units used there are designed for that purpose anyway (it's not like they're dropping in a A19 household LED bulb), so the ballast should be designed for it.

Re:20 year lifespan

[Aguazul2]

Yeah, exactly. Someone somewhere has won a huge contract based on probably unverifiable claims of LED lifespans.

Re:20 year lifespan

[mspohr]

There is no ballast in an LED light.
Fluorescent and sodium vapor lights have ballasts, not LEDs.
All of the traffic lights in my area were switched to LEDs many years ago. I have never seen a single light that wasn't working properly.

Re:20 year lifespan

[necro81]

Because traffic lights go on and off and on and off all day long. They're a terrible idea for LED because the ballasts wear out doing that

LED lighting systems don't have ballasts. True, LEDs require power conditioning (for these applications, it's some sort of switched mode AC/DC converter with constant current output), but those kinds of circuits are highly efficient and robust. LEDs experience essentially zero degradation from being turned on and off repeatedly. All those blinky lights on the front panels of computers, all the flashing indicators on routers and switches, those are all LEDs.

You are probably thinking of fluorescent lamps (tubes and CFLs), for which frequent on/off cycling is indeed a good way to make them die soon. No one makes fluorescent traffic lights precisely for this reason.

Re:20 year lifespan

[jeffb (2.718)]

I think you've got this backwards.

Incandescent bulbs don't thrive in constant-switching environments because there's an inrush of current when the filament is cold. That's why most incandescent bulb failures happen right when you turn the light on. It's less of a problem if you run the bulb below its design voltage, but that drastically reduces efficiency.

LEDs have no corresponding issue. If you're very stupid about the way you build your power supply, that might fail (or even take out the LED), but I don't think that's an issue for traffic lights -- judging from the high-frequency flicker, most of them just run off rectified (and unfiltered) AC. That means they're turning on and off 60 or 120 times a second (50 or 100 in the UK), with no ill effects.

I was going to make a snarky comment about LEDs not using ballasts, but I see that some manufacturers are using that term for LED power-conditioning components. Seems confusing to me, but it's not my field.

Re:20 year lifespan

[Anonymous Coward]

If that's true, then it's a terrible design. Only one of the lights is on at a time, right? So the ballast can run continuously, switching between red/green/yellow as appropriate.

great, so you want to design another single point of failure

Re:20 year lifespan

[Anonymous Coward]

That's actually smart.

Re:20 year lifespan

[avandesande]

LED bulbs require current limiting in the power supply as well or they will burn up- so functionally the have the same requirement as a HPS or fluorescent bulb.

Re:20 year lifespan

[Anonymous Coward]

It says on the box not to run them underwater. What did you think would happen?

Re:20 year lifespan

[pagley]

Well, if you consider a "ballast" as a sort of "power supply", then, yes, there most certainly are "ballasts" in many LED lamps.

LED traffic lights usually have several long-ish strings of LED's (each with a voltage drop of ~2V) wired together so that the drop resistor does not need to dissipate much heat under normal operation - identical in concept to a string of common LED Christmas lights. In the case where there are these long series connected LED strings, it's true that there's no "ballast" per-se.

But, in the case of replacement LED bulbs (40W, 60W, 80W "equivalent" bulbs commonly available now), there is indeed a power supply circuit inside each bulb that regulates the LED current, and some (many, now) provide for phase controlled dimmability too.

LED "flood lights", which would include street lamps, also have a fairly sophisticated power supply in them, particularly important where the power to the lamp housing can be less than ideal, and subjected to large voltage spikes/transients - which can be downright deadly to the light emitting diode junctions in the LED (amongst other things). To get the long lifespan, the power supply circuitry needs to be well designed, built with high quality components, and the LED's themselves need proper thermal management. Not impossible, but not cheap either.

So, in the case of "cheap" LED lights, there isn't necessarily a "ballast". Anything more efficient or higher luminosity than that, there is a "power supply" which would the be LED analogue to a "ballast".

--Brad

Re:20 year lifespan

[mlts]

Nail, head, hit.

LEDs are used for fiber optic transceivers, and flip on and off millions of times a second. If they would end up frying in a short time. Being in IT, I have never encountered failure of a LED (if multiple mode) or laser diode (if single mode, "don't look down fiber with remaining eye") in a network adapter. It either fails immediately, or it works indefinitely. Of course, this isn't 100%, but in my experience, LEDs can handle a lot of switching, far more so than a light bulb filament.

Re:20 year lifespan

[mspohr]

A "power supply" is not a "ballast". Look it up. You may have some point about something but for now, you are only making yourself look ignorant.

Re:20 year lifespan

[mlts]

LEDs are also sensitive to voltage spikes. A light bulb filament can handle somewhat dirty power, but a LED, once it crosses its absolute voltage threshold, is done for. With salt water corrosion, there is always the chance of a short or some electrical glitch which might cause the LEDs to fail.

I've seen some interesting heating element solutions. A small electrical resistance heater that is turned on when the temperature is low is a lot more efficient than a regular light bulb. Of course, when the weather is above freezing, the heater turns off, so that is more energy saved.

Keeping LED lights in traffic signals visible might take little engineering, but it isn't something insanely difficult.

Re:20 year lifespan

[Anonymous Coward]

Not a ballast, a current regulation circuit.
If there isn't any current or voltage regulation circuit, then color me impressed that the LEDs were sturdy enough to have lasted 5 bloody years.

LEDs are way way more solid than most people who've dabbled with lighting technology before their advent can even imagine. Correctly driven and sinked LEDs available now can last for longer than the rest of my lifetime. Almost all of the premature failures are due to poor engineering of the driver and enclosure and the rest are from user-error and because they don't fail instantly and spectacularly, people who do it wrong don't realize it because the thing still works for years even when it's being tortured.

Re:20 year lifespan

[cyberjock1980]

This is precisely correct. LEDs, when properly designed and manufactured, have lifespans that are just phenomenal. I have all LEDs in my house. The ones that I bought that were high grade CREE LEDs I expect to have the rest of my life. Zero failures after 5 years of use so far.

On the other hand, some of the cheaper ones I've bought have often not lasted 2 years. I tried some cheaper ones just to see how well the worked. They often didn't even produce the amount of light they were claiming.

In conclusion you absolutely get what you pay for. And who is going to convince the government to NOT buy from the cheapest bidder. So this will probably be an epic fail since NYC is probably looking at the short term savings vice the long term savings with quality components.

Note that I did not check the actual vendor model to see what brand they are using for this NYC deployment.

Re:20 year lifespan

[kyrsjo]

Sounds like they're doing something wrong. Most of the trafic lights in Oslo* where switched to LED the last 10 years, and I can't remember that ever happening...

* Where it does sometimes get... hmm.. cold. And sometimes we get a few meters of snow too, which stays untill... spring. No polar bears walking around in the streets tough, that's Longyearbyen at Svalbard (island faaaaaaaaar north) :)

Re:20 year lifespan

[Anonymous Coward]

They would if they were sprayed with NeverWet!

Re:20 year lifespan

[pagley]

In the future, I suggest you do a little research of your own before you accuse others being "ignorant". Unfortunately, I fear that bit of advice is nothing more than wasted words on a blunt object such as yourself. But, I welcome you to prove my fears unfounded.

Moving on, let's consult Wikipedia for a reasonable common-man's look at the disputed subject matter... From http://en.wikipedia.org/wiki/Electrical_ballast we have:

"An electrical ballast is a device intended to limit the amount of current in an electric circuit." Seems pretty straightforward.

Applying that to LED lighting, LED's must have their junction current limited. Ideally, for a constant brightness and reasonable life expectation, it should be fairly tightly regulated regulated - usually by a circuit most often referred to as a switch mode power supply, operating in a "current regulation" mode.

Therefore, as I said in my reply, it's proper to refer to it as a "power supply". However, in terms of a "ballast" used in gas discharge lamps, it's performing the same function.

Hence referring to is as an "analogue". Perhaps you need to look that up? Here, I'll help you with that - https://www.google.com/search?q=analogue+definition

Are there any other multi-syllable, "big" words you're having trouble comprehending in either of my replies? I can try to help you with those too if you'd like...

--Brad

Re:20 year lifespan

[WaffleMonster]

A ballast is only needed to limit current for arc-based light sources. So fluorescent lights need them and mercury- or sodium-based street lights need them, but LEDs definitely do not. An LED lamp just needs a power supply and that's about it.

Ballast, power supply.. Tomato tomato.. electronics that drive most LEDs might as well be called ballasts they are still electronic components dealing with relatively high current subject to fatigue and failure. Those gnarly heat fins built into commercial LEDs aint for the leds themselves.

Re:20 year lifespan

[Anonymous Coward]

Tell your city not to buy cheap Chinese crap lights.

Re:20 year lifespan

[ebno-10db]

New Orleans is right on the ocean, so the salt water in the air is more likely to cause corrosion problems.

New Orleans is not unique in being near salt water. If the traffic lights have a problem because of that, then they're badly designed. And yes, I have designed electronic equipment to work in salt fog, etc. Testing for those conditions is common practice. Traffic lights are not some consumer item that you have to baby. They're expensive, and they should be able to take all kinds of conditions and keep working.

Re:20 year lifespan

[Anonymous Coward]

Epic troll.

Re:20 year lifespan

[ebno-10db]

LEDs are also sensitive to voltage spikes.

It would most likely be the power supplies, but either way. There are known techniques for handling all kinds of surges. This is not some piece of consumer garbage. If you can't install it in Central Florida (lightning capital of the world), and have it keep working in any conditions short of a direct lightning strike, then it's a lousy design.

Re:20 year lifespan

[geoskd]

LED bulbs require current limiting in the power supply as well or they will burn up- so functionally the have the same requirement as a HPS or fluorescent bulb.

LEDs function effectively as a diode in a system. When you apply forward voltage, they turn on. For any given voltage, the resistance (and the light output) is constant. Florescent and arc lamps by contrast have a wildly variable resistance at any given voltage, and as such, voltage control cannot be used to effectively protect these lamps from over current. Setting a fixed voltage is easy to do in an efficient manner, setting a fixed current is less simple to do efficiently. In any event, they are worlds and gone different behaviors, so thinking of both types of power supply as a "ballast" is inherently incorrect and potentially harmful.

It should also be noted that the power supply for an arc lamp, or a florescent, have to be capable of producing some pretty high voltages, making them marginally more dangerous. LED drivers by contrast produce the same very low voltage at all times.

Re:20 year lifespan

[kaatochacha]

I guess it depends on the quality of those that were installed. In my area, they replaced many traffic lights with round groups of LEDs, probably a few hundred each for green, yellow and red.
It's VERY common to see one, for some reason usually green, that's missing 10-20 of these lights.

Re:20 year lifespan

[mhotchin]

I suspect that the snow build up is highly termperature dependent - very cold snow won't stick to anything. In N.A., many cities have winters and snowfalls very close to freezing temperatures, which is the perfect temp for sticky snow.

Re:20 year lifespan

[c 0xFF]

For high-power LEDs, the failure mode is generally a reduction in light output, not a complete failure of the LED itself. If there is a complete failure, the power supply is most likely at fault. Again, the switching isn't the problem, but more likely exposure to the elements affecting the circuitry.

If I were to guess, New Orleans found itself with a bad batch of electronics or housings, and budget constraints forced the installation of cheaper-in-the-short-term traditional bulbs.

Re:20 year lifespan

[kyrsjo]

It's not always cold:
http://www.yr.no/place/Norway/Oslo/Oslo/Oslo/statistics.html
Sorry if it's in C not F, but in general: 0C is the freezing temp of water, 20 is comfortable, 30 is really hot, -5 is normal cold, -10 is quite cold, and -20 is freeze-your-balls-off-cold. Luckilly, that doesn't happen to often.

Generally the winter weather in the city (which is by the sea, or at least a small fjord) is quite variable. The proximity to water also means that it *may* be humid - and humidity amplifies the feeling of coldness. -20 and dry is really prefferable to -1 when its snowing wet snow (melting on the ground, freezing overnight... Makes it interesting to live in the university student village where most of the exchange students also live...), especially if there is any wind chill...

So yeah, we also do get sticky snow, especially at the ends of the seasons. The frequent temperature cycles also makes for tons of ice. Still, this isn't a problem for trafic lights - I guess it's down to the design. All trafic light lamps have a "baseball cap" sticking out over them, which is rounded and smooth on the top, meaning that wet snow just slips off. Also, there is no need for collimator "blinds" to make it visible in strong sunlight. An finally, there has been a trend for the last ~30 years or so (as in most of Europe) to build roundabouts wherever there is space for one.

Re:20 year lifespan

[Bayoudegradeable]

Ahhh. thanks for the info. Never thought about the on off thing. Ballasts, huh.

Re:20 year lifespan

[Bayoudegradeable]

be nice : )

Re:20 year lifespan

[Agripa]

Discharge lamps have a negative resistance characteristic so they require ballasts to limit their current. That can be as simple as a resistor but is almost always a reactive element for low power dissipation. Electronic gas discharge ballasts are high voltage compliant current sources. Laser tubes often include a ballast resistor.

LEDs have a low dynamic resistance combined with significant variation in forward voltage drop as well as a negative voltage temperature coefficient all of which combine to create much the same problem as well as current sharing issues if LEDs are operated in parallel. Without ballasting, parallel LEDs are subject to thermal runaway.

Both use reactive ballasts in one form or another that produce a current output. Gas discharge lamps operate at high voltages and require even higher voltage starting. LEDs require additional provisions for current sharing and failsafe operations of long strings.

For traffic signals LEDs are obviously superior except maybe in harsh conditions but I am not convinced that is so for street lighting where they are less efficient. Their more complex ballasting and package requirements lead to poor reliability compared to the individual LEDs themselves.

Re:20 year lifespan

[Agripa]

LED lighting systems don't have ballasts. True, LEDs require power conditioning (for these applications, it's some sort of switched mode AC/DC converter with constant current output), but those kinds of circuits are highly efficient and robust.

"Some sort of switched mode AC/DC converter with a constant current output" is a ballast. That is exactly what you find in an electronic ballasts designed for discharge lamps. In consumer based lighting, the electronic ballast used for either discharge lighting or LEDs is the least reliable part.

Re:20 year lifespan

[Agripa]

I get about 6 months on Edison base compact fluorescent or LED lamps because of summer lightning storms. Incandescent are much more economical for me.

Re:20 year lifespan

[thegarbz]

For any given voltage, the resistance (and the light output) is constant. Florescent and arc lamps by contrast have a wildly variable resistance at any given voltage, and as such, voltage control cannot be used to effectively protect these lamps from over current. Setting a fixed voltage is easy to do in an efficient manner, setting a fixed current is less simple to do efficiently.

You're missing one critical thing. The current / voltage curve on a diode system is almost vertical around the forward bias point and the reverse breakdown point. Just because you can set AN LED at the correct voltage and then run it via voltage control doesn't mean you should. Manufacturing variances mean an identical voltage applied to two different LEDs mean a wildly different current is consumed, one may light dimly the other may blow. Same goes for variances throughout their life.

Voltage control on LEDs is a very very stupid thing to do if you care at all about their life. There's a reason every household LED down light has a constant current driver behind it.

Re:20 year lifespan

[thegarbz]

All the LEDs in your typical computer are resistance limited. Typical high power LEDs make this totally impractical due to resistance losses. They are typically driven by a constant current supply, a switchmode supply, a supply that if you don't take care in designing or buy decent reactive components for may have a limited life if you start hitting them with powercycles / brownouts.

LEDs only ever really die if the designer was stupid or the driving circuit breaks and overdrives them. Same thing applies to CFLs which as stats will show die quite frequently.

Re:20 year lifespan

[Anonymous Coward]

Led lights do NOT last 20 years - I call bullshit, and bet the inverters needed to drop the voltage fail in 5-10 years tops. Often the inverter will drift blowing up the leds as well. Street lamps in full sunlight/heat, then snow suffer temperature cycling, again reducing lifetimes. Finally, a good bet that they will be made in China or the lighting bit will be. If they overdrive the leds, lifetime will be shorter.

By the same people who told you CFL's last 5-8 years. Last time I looked led's output drops a lot in cold temperatures.

Re:20 year lifespan

[vandamme]

We (Rome NY) replaced them all many years ago, and I never see them replacing them, and one or two "sick" ones.

Re:20 year lifespan

[vandamme]

A ballast creates a current source from the mains voltage source, balancing out the negative impedance characteristic of a fluorescent bulb or LED. It can take the form of a reactive circuit element such as an inductor, capacitor, or dissipative resistor, or an electronic circuit which feeds the proper constant current. To make a constant voltage power supply into a constant current power supply, you change the feedback loop.

Thank you for your opinion, though.

Re:20 year lifespan

[vandamme]

I see a lot of LED flashlights that don't use any kind of current limiter, relying on the small IR drop through the LEDS and the battery resistance to avoid runaway. Well, I suppose 100 hours is a long life for a flashlight or disposable toy light.

Re:20 year lifespan

[thegarbz]

These torches rely on two things that don't apply to larger scale lighting installation:
a) Resistance either external or within the battery is used and sized in such a way that the maximum draw won't exceed the max operating point of the LED.
b) External resistance is used for current limiting.

This creates several problems for running LEDs:
- The torch is only at maximum brightness briefly (circuit designed for freshly charged batteries which will likely drop 0.1V in the first minute or two of use.
- LEDs don't run at max brightness / efficiency during normal use.
- Power is lost in resistance and is wasted as heat.

Now this is actually not a problem in torches, and is how small LEDs in computers / phones are driven because there's no efficiency requirements. However if you scale up the size of the LED things start looking very different. Ignoring the fact a small blip in the powersupply could ruin an LED, if you have a 20W LED to control your voltage your source resistance may suddenly waste a lot of power. But you still run the risk of a slight disturbance burning up your LEDs.

The most efficient way to drive an LED is also the one that lets it shine brightest and isn't actually any more difficult to implement in a fixed install. If you're not running from batteries chances are you're running from 240V at which point you need a powersupply to convert the 240V to something more LED friendly. In this case instead of the powersupply targeting a specific voltage and letting the load draw as much current as needed to keep that voltage stable, the powersupply provides a fixed current and varies the voltage to ensure that the load current doesn't change.

Incidentally if your cheap torch runs from a single AA then this method needs to be used as well since a single AA can't supply the required voltage to light the LED. The entire powersupply circuit can be about the size of your thumbnail.

Re:20 year lifespan

[dewrox]

It sounds like they chose to implement sub-standard LED's because I know from experience that the LED lights that are used on Semi Tractor Trailers are quite visible from well over 5 miles with a foot of snow covering them on the tail end of the trailer... all in -25 to -35 degrees Celsius. This would lead me and most other people that have practical experience in this area... that they used cheap LED's for the street lights if the light is not visible under a little snow. And yes this was observed in Alberta, Canada during a snow storm. It was when I finally caught up to the truck that I was able to see how much snow was covering the lights.

Re:20 year lifespan

[lsatenstein]

There is no ballast in an LED light.
Fluorescent and sodium vapor lights have ballasts, not LEDs.
All of the traffic lights in my area were switched to LEDs many years ago. I have never seen a single light that wasn't working properly.

My LED bulbs have a miniature power transformer in the base of the lamp, so that the voltage is stepped down, and with the high resistance windings, the ballast effect is provided. The open circuit voltage is 14, operational voltage 11-12 volts. Some LEDS can work with dimmers, other cannot.

Re:20 year lifespan

[mspohr]

Please learn something before you go spouting ignorance.
Start with Wikipedia: "a ballast limits the current in an electric circuit".
You are talking about voltage, ballasts limit current. If you don't know the difference between voltage and current, do a little research.
LEDs use voltage regulators.
Florescent and gas discharge lights use a ballast to limit current.
They are different. Voltage and current... it's not complicated but it does require some education.
You cannot just wave your arms and mumble a lot of words and pretend they are the same.

Re:20 year lifespan

[lsatenstein]

A ballast can work with low voltage as well, We had saturated cores, so that as the draw went up, the efficiency of the transformer went down. I worked with this type of equipment in industry. It is not called a ballast, but it does regulate output power and voltage, and waveform.

So I wave my arms, and say mumble jumble, soft iron core transformers come back to do your thing.

Re:20 year lifespan

[pagley]

You cannot just wave your arms and mumble a lot of words and pretend they are the same.

I think you should be heeding some of your own advice there, buddy. Normally, I don't feed trolls, but I'm feeling particularly ornery today, so here goes...

So you're now quoting the very same layman's definition of a "ballast" from Wikipedia I referenced in our previous discussion here to belittle someone else for not being "educated"? This, coming from the same troll who was unable/unwilling to comprehend the fact that an offline LED power supply (whether a simple, linear deign, or a more complicated SMPS type) was analogous to a gas discharge lamp ballast... That's rich.

LED's require their junction current to be limited, period. However, when the power source cannot be configured to supply "constant current", and is a "constant voltage" type, a series resistor is inserted in circuit, with it's value selected based on desired junction current, source voltage, and junction voltage. That series resistance (whether supplied from winding resistance, external resistor, or a combination of both) is known as... ballast resistance.

Rather than continuing to cultivate your own astonishing ignorance on subject matters you obviously know less than nothing about, I'd suggest you search that empty head of yours and try to locate two brain cells to rub together, educate yourself before you speak/type, and try to form a coherent thought. I know that may be a stretch for you, but give it a try - you'd be a better person for it.

And, if you aren't willing to actually try, then shut your trap.

--Brad

How many Mayors

[mdsolar]

Does it take to change a light bulb?

Re:How many Mayors

[Anonymous Coward]

Two. One to send an RFP to SAIC, and the other to mix the martinis.

Re:How many Mayors

[Valdrax]

Apparently, 1/250,000th of one.

Re:How many Mayors

[vandamme]

I'd worry about some of the dimbulbs running for election.

Cue the Wiener jokes.

Many of those bulbs due for replacement anyway

[drnb]

Yes, there is a savings, but how much is it going to cost NY taxpayers up front ? Would a better strategy be to replace the sodium lights with LED style lights, as they wear out?

Yes, there is a savings, but how much is it going to cost NY taxpayers up front ?

It looks like a 4 year program and the incandescents last about 7 years. So many of those bulbs will be due for replacement anyway.

Re:Many of those bulbs due for replacement anyway

[avandesande]

Keep in mind to that HPS bulbs, ballasts etc. are pretty expensive- my guess is an installation costs at least a couple hundred each. There aren't any downside to the retrofit that I can think of.

Re:Many of those bulbs due for replacement anyway

[djlemma]

They are replacing High Pressure Sodium lamps, which are not incandescent. The funny thing is, by the standard measure of efficiency used in the industry, the new street lamps probably will be LESS efficient than the old ones. HPS lamps can get above 100 lumens per watt pretty easily, and low pressure sodiums can even get up to 200 lumens per watt. They've been able to get efficiency like that in labs for LED's, but for production fixtures it's not very common.

Of course, LED's often win out in real-world comparisons, because all the lumens are more efficiently directed where they need to go. Still, to get that much brightness, it's going to cost quite a lot of money.

Re:Many of those bulbs due for replacement anyway

[evilviper]

HPS lamps can get above 100 lumens per watt pretty easily, and low pressure sodiums can even get up to 200 lumens per watt. They've been able to get efficiency like that in labs for LED's, but for production fixtures it's not very common.

Some of the cheapest 120v LED bulbs out there (like the ones I just bought) are over 80 lumens per watt, so an easy match for HPS.

Production LEDs aren't as efficient as LPS, but you get vastly better color rendering in the trade, which has been the reason for many cities to use far, far less efficient street lights in the past, so LEDs are still a big upgrade by most measures.

Still, to get that much brightness, it's going to cost quite a lot of money.

Perhaps there's something corrupt going on in the street light market, I don't know. But in the consumer market, the expense of the ballast and bulb drives up the cost of HPS and LPS fixtures to about twice as much as consumer LEDs of the same output.

They've already done this here

[Anonymous Coward]

in Buenos Aires, lots of streets have already been replaced, but we still have a lot to go.

The effect is quite strange. The cold-white light reflects off the pavement in a spot (as opposed to the diffused old orange-yellow bulbs), making streets look permanently wet-ish.

Doesn't look bad, but it's different.

Not Incandescent

[Anonymous Coward]

No reasonable community has used incandescent street lights for decades except for certain historic districts. High-intensity discharge lamps standard. The summary clearly has no idea what it's talking about.

I can but hope

[Anonymous Coward]

I really hope they will use models that reduce light polution. Yes, it seems futile in such a city, but if you're replacing wholesale anyway, might as well add this. In fact, a little care could reduce glare and make lighting safer that way, with a nice side benefit of less wasting light upward. On that scale even a few percent light not wasted adds tangible savings.

Re:I can but hope

[mmclure]

The picture in TFA seems to show a modern full cutoff fixture (i.e. a fixture designed so that the light only projects down and sideways, not upward.) Hopefully they will resist the temptation of thinking "well, LED lights use less energy so lets put brighter LEDs in!" that has happened in other communities that replace their lighting with LEDs.

Stick with sodium

[mdsolar]

Near observatories to cut down on light pollution. LEDs are too broadband.

Re:Stick with sodium

[drnb]

Near observatories to cut down on light pollution. LEDs are too broadband.

In the digital era astronomers no longer need to be near their telescope. We can put the telescopes in more remote locations, or ideally in space.

Re:Stick with sodium

[fuzzyfuzzyfungus]

Near observatories to cut down on light pollution. LEDs are too broadband.

On the plus side, if somebody is thinking about installing LEDs, that is (sometimes) a sign that light fixtures that have been, well, fixtures, for decades, sometimes quite a few of them, are getting their first re-evaluation in quite some time.

It only helps if somebody pushes at the correct time; but if the fixtures are being reevaluated in anything resembling a serious way, that's your best chance to get action on things like fixtures that point upward, ill-designed fixtures that don't target their output very well, and all the various other dubious lighting decisions that help add up to light pollution.

It's unlikely to be perfect; but LEDs (being costly; but easy to aim fairly tightly, as well as very good at doing accent work (say, lighting a set of stairs with small lamps set just above the steps, rather than one big bulb-on-a-stick pointed in the direction of the stairs and cranked to 11), do encourage more efficient targeting in a way that big, cheap, one-size-fits-all bulbs don't.

Re:Stick with sodium

[Joce640k]

What about amateur astronomers?

Amateur astronomers actually make a *lot* of the discoveries and do a lot of the photography.

Re:Stick with sodium

[khallow]

And are you going to pay to put and maintain telescopes in those places? The problem here is that we're discarding existing infrastructure without adequately building replacement infrastructure. The US government is notorious for underfunding such things.

Re:Stick with sodium

[SJHillman]

From the few astronomers I've talked to, streetlamp design affects light pollution as much or more than what kind of bulb it's using. Many streetlamps, especially older ones, shine in every direction including up. Lights with a hood to reflect all of the light towards the ground are better for both energy consumption and light pollution.

Re:Stick with sodium

[mdsolar]

Yes, directing light to where it is needed is important as well.

Re:Stick with sodium

[jratcliffe]

Not a lot of astronomy going on in NYC.

Re:Stick with sodium

[Anonymous Coward]

In the digital era astronomers no longer need to be near their telescope. We can put the telescopes in more remote locations, or ideally in space.

Yeah, who the fuck needs to see the sky anymore. I bet we will have lots of kids wandering about space and science in general, when they can't see more than 1 star in the sky.

Re:Stick with sodium

[Ichijo]

Also, because LEDs come to full brightness faster than sodium lamps, it raises the possibility of using motion sensors to turn them on only when needed, reducing light pollution and energy usage.

Meanwhile, there's some debate about whether street lights reduce crime. They create dark pockets where attackers can hide, and the illumination they provide helps burglars see what they're doing.

Re:Stick with sodium

[AK Marc]

LEDs are more directional. Direction matters more. Always light billboards from the top down, never the bottom up. Small changes can eliminate most of the light pollution.

Re:Stick with sodium

[SJHillman]

Even small cities can affect light pollution from dozens of miles outside of the city. I grew up about 20 miles from a city of 60,000 and it gave off enough light pollution to effectively blot out much of the southern sky near the horizon. I imagine NYC's light pollution reaches much further.

Re:Stick with sodium

[Ichijo]

And one more potential issue with white LEDs is that white light with a high color temperature disrupts your sleep.

Re:Stick with sodium

[drnb]

What about amateur astronomers?

Even more telescopes in space. We're supposed to be commercializing space right? Rentable telescopes. You rent it for an hour, you point it in the direction you want, gimbals permitting.

Amateur astronomers actually make a *lot* of the discoveries and do a lot of the photography.

How many amateurs make discoveries while in NYC? Or are they driving to the Poconos or Catskills?

Re:Stick with sodium

[edxwelch]

they also effect the migrating patterns of badgers, or something

Re:Stick with sodium

[umafuckit]

I'm an amateur astronomer and it's become pretty clear to me that we're fucked with these bright white LEDs. Light pollution is already very, very, bad and these bright, broad-band, light sources are effectively going to finish the job and destroy the night completely for locations within 100 or so miles of any urban area that uses them. In areas of high population density it will become impossible to see the night sky without traveling vast distances. Lighting levels have been increasing at a far greater rate that population is increasing. There's a constant push to make things brighter but little thought as to lighting more thoughtfully or efficiently (by which I mean lighting only what needs to be lighted and not over-lighting).

Re:Stick with sodium

[cdrudge]

They create dark pockets where attackers can hide

So the street light creates darkness? If anything, I'd say the streetlight reduces the size of dark pockets.

Re:Stick with sodium

[Anonymous Coward]

Statistics suggest that with enough lamps, crime goes down. Having only a few lit streets only means less crime under the lamps.

As for dark pockets: without lights, it is dark everywhere after sundown.

Re:Stick with sodium

[SJHillman]

If you're walking under streetlights, obstacles create shadows (pockets of darkness). Your eyes are adjusted to the lights, so you can't see into these pockets, but muggers in the shadows sure can see you.

Re:Stick with sodium

[Alomex]

But they are more directional, counteracting the broadband effect. Plus are there really any observatories near NYC? I think you are just complaining for complaining sake.

Re:Stick with sodium

[Anonymous Coward]

(pedantic) It's not the LEDs, but the phosphor on the white LEDs that are broadband.

Re:Stick with sodium

[Anonymous Coward]

What range of lot are you talking about? The 0.074% discovered by amateurs in the last 3 decades that have contributed to the field, including those that take real research data as their base, or is the 0.074% some higher value that you believe? Taking photos doesn't count, slow exposure time lapse is rarely gets passed reddit shovel vanity images, and none make it into publications of note.

Re:Stick with sodium

[Anonymous Coward]

That's an incorrect assumption. Here in Orlando they are doing the same LED retrofit, and one of the advantages is actually reduced light pollution.

The LED fixtures are substantially more directional than the existing fixtures uselessly send up to half their light up into the atmosphere.

The City of Orlando is not only saving money on electricity and maintenance, they are making the night sky darker in the process.

Win Win.

Re:Stick with sodium

[Anonymous Coward]

LEDs are more directional. Direction matters more.

Depends. For low pressure sodium, it has a very narrow light emission spectrum. There are notch filters available that will filter out the entire low pressure sodium spectrum leaving the rest alone. And then the sky is dark once more when looking through a telescope.

LEDs, on the other hand, completely screw this up by emitting broad spectrum. You can't filter them efficiently. Game over.

  http://palomarskies.blogspot.com/2009/05/why-astronomers-love-low-pressure.html

So sure, direction matter if all you care about is general brightness of the sky. But for people with filters trying to live with light pollution, LEDs is game over.

This is aside the health effects that light pollution has and the low-pressure sodium lights are more efficient than LEDs (and don't disturb the circadian rhythm).
      http://www.eoearth.org/view/article/153402/

Re:Stick with sodium

[c 0xFF]

LEDs are more directional, meaning that more light goes down, instead of up. Do you think that would help any?

Re:Stick with sodium

[Anonymous Coward]

I'm an amateur astronomer and it's become pretty clear to me that we're fucked with these bright white LEDs. Light pollution is already very, very, bad and these bright, broad-band, light sources are effectively going to finish the job and destroy the night completely for locations within 100 or so miles of any urban area that uses them. In areas of high population density it will become impossible to see the night sky without traveling vast distances.

Lighting levels have been increasing at a far greater rate that population is increasing. There's a constant push to make things brighter but little thought as to lighting more thoughtfully or efficiently (by which I mean lighting only what needs to be lighted and not over-lighting).

Light pollution is almost non-existent if the atmosphere is not already heavily polluted. If you look up at the sky at night and don't see something like this: http://www.eso.org/public/archives/images/screen/epod-cc-rf18284.jpg
The problem is not light pollution, it's air pollution which is scattering the light back down. Fix your smog problems and all the LED's in the world aren't going to cause you any major issues.

Re:Stick with sodium

[umafuckit]

LEDs are more directional, meaning that more light goes down, instead of up. Do you think that would help any?

This was the hope, yes. Unfortunately it's not working out that way. Mainly this is because these LEDs are *so* much brighter that the directionality advantage is lost since we're seeing more light reflecting off the ground and into the sky. Also, some direct light likely does creep into the sky from the fitting. Secondly, the directionality advantage assumes the fittings are attached correctly (sometimes they're not). In addition, it's now becoming possible for home users to buy LED "globes" with LEDs pointing in all directions. Thirdly, LEDs are broad-band and so they hit the peak absoption of your rods, which inhibits dark adaptation and masks out the objects you're trying to see. So white LEDs are not only churning out more light, but it's "worse" light.

Finally, we do have narrow-band "light pollution filters" for visual astronomy. These work well for emission nebulae which emit fluorescence at particular wavelengths. The filters let through these wavelengths and cut out the red of sodium lights. However, with more and more "white" light pollution, the effectiveness of the filters will go down considerably. Furthermore, no filter helps on galaxies because these are composed of stars and so are broad-band emitters.

The night sky is a massive part of our natural heritage and an important gateway into science. It's a pity we're wiping it out, because the pretty pictures aren't a patch on seeing things for real (even if what the human eye can see is so much less). I do astronomy outreach and regularly have adults babbling like kids when they're at the telescope. The observatory I attend will be dead in 10 or 20 years because of light pollution, which each year gets visibly worse.

Re:Stick with sodium

[umafuckit]

It's unlikely to be perfect; but LEDs (being costly; but easy to aim fairly tightly, as well as very good at doing accent work (say, lighting a set of stairs with small lamps set just above the steps, rather than one big bulb-on-a-stick pointed in the direction of the stairs and cranked to 11), do encourage more efficient targeting in a way that big, cheap, one-size-fits-all bulbs don't.

Except that most people think a big bright light is better and LEDs already can be purchased as fittings that encourage this use. It's cheaper to set up one bright light that floods everything than lots of small integrated accent lighting. The fact that the accent lighting will be far, far, nicer, and cheaper to run doesn't enter people's heads. Also, there's a perception that illuminating the shit out of something will reduce crime. It is unclear whether this is really case. The only thing that is clear is that public lighting increases people's subjective feeling of personal safety.

Re:Stick with sodium

[umafuckit]

Exactly. A similar effect occurs if you're sitting around an outdoor table at night and the only form of illumination is a bright lamp in the middle of the table. You can't see the faces of most people sitting around the table when you do that. However, if the lamp is placed further away, illuminating the scene less but more uniformly, then the problem goes away.

Re:Stick with sodium

[umafuckit]

No, but the skyglow from NYC spreads at least a hundred miles. I live 40 miles away and its destroyed amateur astronomy for me.

Re:Stick with sodium

[AK Marc]

Then work with LED manufacturers to make multi-spectrum white (i.e. a mix of three colors, rather than a broad spectrum). I thought there were already complaints about white LEDs being non-broad-spectrum by the health nuts? Funny when two people hate the same thing for mutually exclusive reasons. Your bane is incandescent. They are full-spectrum.

Re:Stick with sodium

[evilviper]

Holy run-on sentences, Batman!

What the hell did periods ever do to you, to make you hate them so very much?

Re:Stick with sodium

[Maury Markowitz]

> Near observatories to cut down on light pollution

The vast majority of useful astronomy takes place in far remote locations like Atacama and the top of Mauna Loa. Complains from telescopes near urban areas, like the DDO and Mount Wilson are victims of urban growth, no light source is going to fix their problems.

Lots of complaints from the amateur astronomers, sure, but they can move their equipment in the back of the car.

Re:Stick with sodium

[Maury Markowitz]

> In areas of high population density it will become impossible to see the night sky

Impossible eh? We just went to a star party at a marina, under parking lot lamps. Saw the coal sack and Orion just fine. Maybe you need new eyes?

> without traveling vast distances

Good thing we invented cars then.

> The night sky is a massive part of our natural heritage

And there's more of it all the time, a side effect of people moving out of the country into the cities.

Don't be lazy, the journey is the reward.

Re:Stick with sodium

[Maury Markowitz]

> Meanwhile, there's some debate [newyorker.com] about whether street lights reduce crime.

No there isn't. There's a single statement in the entire multi-page article about crime, and it specifically refers to security lighting. There is nothing in the article that could be considered "debate", and certainly no hint of anything as generalized as you imply here. It's the New Yorker, so don't expect facts anyway.

With the power of Google at my fingertips, I went looking for a credible meta-study and found one almost instantly:

http://www.crim.cam.ac.uk/people/academic_research/david_farrington/light.pdf

I quote from the introduction, "Results of this review indicate that improved street lighting significantly reduces crime."

What *is* interesting from the results is that "The review also found that nighttime crimes did not decrease more than daytime crimes." This means that the use of lighting is changing patterns of crime not through direct observation, but due to increased social cohesiveness.

That said, all of this ignores the overarching fact that crime is going down, all the time, faster than anyone imagined. Whatever speculative effect lighting *might* have, I guarantee the former is orders of magnitude greater.

>They create dark pockets where attackers can hide:

Oh geez, that is the dumbest argument I've ever read. The lack of street lights creates one gigantic "dark pocket", so obviously that's worse, right?

Crime goes down in lit areas. Period.

> and the illumination they provide helps burglars see what they're doing.

This is the second dumbest: the same person just argued the criminals use the dark areas.

Re:Stick with sodium

[Maury Markowitz]

> Your eyes are adjusted to the lights, so you can't see into these pockets, but muggers in the shadows sure can see you.

Which, I guess, is why lit areas have significantly less crime, and that the crime rates keep going down. Right?

Re:Stick with sodium

[umafuckit]

I said it *will* become impossible, not that it is now. Besides, you'll probably always be able to detect the very brightest DSOs; I've found M81 and M82 from the London suburbs under terrible conditions. However, they're merely "detectable" and appear as only a shadow of what they actually look like under dark skies. Very impressive objects, such as M33, aren't visible at all under those conditions. The night sky itself looks shit to the naked eye. Astronomy under these conditions gets boring pretty quickly and conditions are only getting worse.

I don't know what your marina was like, but you describe only a local source of light pollution. In my experience, the sky glow from a big city is far worse than observing next to a few lamp posts. If you stick a towel over your head when at the eyepiece and allow yourself to dark adapt. A great deal can then be seen if you're not too near a big city.

Your sarcastic comment about cars misses the point. I do currently travel with my telescope to dark skies, but to get to mag. 6 I have drive 3.5 hours and spend the night. I don't often have the chance to do that because it's effectively a 2 day commitment and it has to be juggled alongside work, weather, and the phase of the moon. Usually I can only do the 1.5 hour trip to ~mag 5 skies. Then there's still 30 minutes of so of set up and then tear-down. I make the investment: I'm not lazy. Fuck, I've driven 3000 miles for a star party in the past. But I rarely get the time for that and the journeys are getting longer for the same reward due to the spreading lights.

I don't know what you mean by "there's more of it all the time." We're losing many of the dark skies (even in rural and semi-rural areas) we currently have *despite* any depopulation of these areas to the city.

Re:Stick with sodium

[SJHillman]

I never said it increased crime in general, I just said it provides an easier opportunity for crime in locations that don't have uniform lighting.

why?

[Anonymous Coward]

Are LED really that much more efficient than Sodium vapor it is worth going through and replacing perfectly good bulbs?

Replacing as they fail makes more sense, compare up front cost to lifetime savings and do or dont. Going though and doing an extra change cycle (labor) to replace bulbs that are still good doesn't seem to make sense.

Re:why?

[Amouth]

Actually this is a simple math problem.

Two options:
Replace all at once
Replace as they burn out

Either way i have to physically replace each bulb.

It is more cost efficient to replace them all at once in a sequential pattern, rather than one at a time randomly, Thats because the cost to replace is the same, but i'm minimizing my travel distance & times as i'm going dispatch->pole->pole->dispatch rather than dispatch->pole-dispatch->pole->dispatch. You would be surprised but travel times are normally the highest impacting item when it comes to wrench time measurements. Also to add to it, if i wait and replace as they fail i'm paying X for electricity over that time, where if i replace it now i pay Y which is lower than X. The power savings is a fringe benefit compared to labor, but non the less it is factored in.

The biggest question that comes to mind for this type of decision is the time value of money. I can spend X now or X+1 from Now till then. which one is lower cost overall between now and then isn't always a straightforward answer.

Re:why?

[Richy_T]

It's actually a somewhat more complex statistical problem and will include, at the very least, the profile of how bulbs wear out.

I would imagine it would be best to replace one-at a time to somewhere near (or possibly beyond) the mean lifetime and then replace all at once (and even then, you may or may not want to replace ones that have already been replaced recently).

Re:why?

[kyrsjo]

They might also have to move the poles around / add more poles. So it migth not be possible to do it on a pole-by-pole basis.

Nothing

[mdsolar]

Cities can and do borrow to pay for long term investments. If NYC issues bonds, then this is the sort of thing they support. The taxpayers see saving all along the way. If the bond rate is below inflation, they see even more.

LEDs break?

[Anonymous Coward]

How do LEDs break? I've never seen it happen in any of the LEDs I've owned.

Re:LEDs break?

[rubycodez]

LED have crystalline material grown on a substrate. one cause of failure is thermal stress, causing cracks to grow from the boundary. Another main cause of failure is structural defects in the crystal creating more paths for leakage current carriers. Those cause even more structural defects or make existing ones grow.

New city motto

[Anonymous Coward]

NEW YORK CITY: Leaping boldly into the Year 2000- Today!

At least it's not CFL

[Cammi]

At least he didn't go CFL. Those suckers have a life span of 5 months. Haven't tried LEDs yet, but all what I have seen so far is that they are about 5% brightness of CFL and Regular bulbs.

Re:At least it's not CFL

[__aarzwb9394]

CFL. Those suckers have a life span of 5 months

They must be diffrerent from the CFL bulbs that have lasted for over 2 years in my house.

How is it possible that a tech forum is full of change-phobic luddites?

Re:At least it's not CFL

[jratcliffe]

At least he didn't go CFL. Those suckers have a life span of 5 months. Haven't tried LEDs yet, but all what I have seen so far is that they are about 5% brightness of CFL and Regular bulbs.

Yes, if you compare a LED bulb with x lumens with an incandescent of 20x lumens.

Re:At least it's not CFL

[foobar bazbot]

Yes, it's true! Street lights use exactly the same bulbs as your home lighting fixtures, so we can expect they'll use exactly the same dollar-store 3W LED bulbs that you've seen. Or maybe, since streetlights actually use 200W-1000W HID bulbs, the LED replacement they're talking about will have virtually nothing in common with the LED bulbs you're talking about, so you're really talking without a clue...

Re:At least it's not CFL

[Anonymous Coward]

Your power must be really flakey or something. I have CFLs in my basement that are at least 10 years old and still working fine (well, as fine as they did when new, anyway - slower to startup than modern ones, and more of a yellow color). I also have a few LEDs from Costco which are great 60w replacements; they actually seem a little brighter. We also just put in a bunch of MR16 LEDs in track lighting - they are easily brighter than the 30w bulbs they replaced, at only 6w, and the color is really nice. Those suckers were expensive though (Sylvanias with a 5 year warranty).

Re:At least it's not CFL

[Lonewolf666]

Depends on the vendor. I had pretty good quality from Osram and abysmal from Megaman (both CFL brands in Germany).
From Osram, 10 years or more of lifetime seems normal (small sample size here, but it points in that direction).
From Megaman, I bought four CFLs and three of those broke after a few months. Not a vendor that will get any more business from me...

Re:At least it's not CFL

[Richy_T]

There are now some decent LEDs out there. You'll pay for them though.

Haven't found a CFL worth a damn though. I think I have one left still working from the time I decided to go on a CFL-installing spree.

Re:At least it's not CFL

[Anonymous Coward]

Burned by the early adopter penalty too many times.

Sent from my Newton.

Re:At least it's not CFL

[iroll]

Well, maybe you just haven't seen that much, then. I've got a mixture of CFLs, LEDs, and incandescents in my house, and they all work just fine, with lifespans that are long enough that I can't remember when I changed any particular bulb. While lighting is an insignificant cost compared to running my air conditioner, I appreciate that the modern bulbs use significantly less juice, and they don't heat the place up as much.

So there you go, an anecdote for an anecdote.

Re:At least it's not CFL

[ebh]

I have lots of CFLs and their lifespan varies widely. I've had some going for almost five years now, and others didn't last five months, all indoors and on roughly the same duty cycle. I don't have too many LED bulbs yet, the oldest maybe 18 months, but none of them have failed yet. I guess LEDs don't have as much "infant mortality". OTOH, we have some incandescents that were here when we bought this house 17 years ago and are still humming along.

It'd be nice if we could get a nonbiased study of lifespans, changes in output, etc., based on a decent sample size and not connected with any manufacturer.

Re:At least it's not CFL

[Cammi]

Exactly. We switched from the regular bulbs to CFL to save money ... but in the long run, CFLs cost more at our house due to their low survival rate. The cost of replacing them is more than what they save on the power bill.

Re:At least it's not CFL

[Anonymous Coward]

LEDs work out to be brighter than either incondescents or CFLs, but they tend to have a much narrower light cone (although Philips new replacements for A19 bulbs are supposedly pretty good), so they are good replacements if you have pin spots or overhead cans. The other of drawbacks of LEDs are that they need to stay relatively cool to operate, so they don't yet work out for some high lumen applications, and that they don't work well with pulse modulated dimmers (usually denoted by a warning "use with tungsten loads only") --you get a nice strobe affect going below about the 85% setting. However, they do work fine with voltage modulated dimmers (unlike CFLs --even the supposedly dimmable ones), except that "off" is usually at around the 5% setting on the dimmer.

Re:At least it's not CFL

[Anonymous Coward]

Yup. I'm a change-phobic Luddite because I report the truth. No, I am not the OP, but I have had a similar experience with CFL in two houses. Rooms where lights are turned on and off frequently like bathrooms and I have CFL bulbs that I replace more often than incandescents. Rooms like my living room where they stay on for hours and then left off for hours, they last for years.

Perhaps you might want to not assume that people who have different experiences than you are not idiots or deficient.

-- MyLongNickName

Re:At least it's not CFL

[mlts]

I've had decent luck with CFLs, with the only downside being that one has to be careful due to the small amount of mercury in them if they get broken (so I would try to find enclosed bulbs for safety reasons.)

The latest generation of LED bulbs is heads and shoulders above CFLs in almost every respect that I tossed all the glass tube lights. If dropped, it might dent the heat sink, or at worst dent the bulb portion, no broken glass. The life expectancy is a lot longer than CFLs, and they do not use much electricity, especially compared to incandescent bulbs. Of course, color temperature matters a lot. The cheapies have the undead-looking blue-white while the more expensive ones have a warmer white.

LEDs are also a must when RV-ing. LED bulbs use 1/5 to 1/7 as much electricity as incandescent bulbs, which make a big difference when boondocking in the middle of nowhere.

Re:At least it's not CFL

[Anonymous Coward]

CFL lifetime is usually a function of number of cycles done or power in your house. For most bulbs the ballast goes before the bulb does. My parents have bulbs that I installed still going from when I was a kid (20 years ago). If you have a power fluctuation problem they will burn fast.

Most of the bulbs I have installed are going on 5 years now. I have a box of 30 that the power company gave me. I have used 8. The other incandescents I only replace when that room/hall gets more use. The ones I use every day were replaced years ago.

I am waiting on cost to come down for LED to switch out. I have my eye on some overhead halogen lamps that make the room quite toasty when they are turned on.

Re:At least it's not CFL

[bunratty]

Since I replaced most of the bulbs in my house to CFLs, I'm changing light bulbs far less often. They last far longer than incandescents in my experience, exactly as claimed on the box. I've also bought several 65W LED floodlights, and they produce light just as good as incandescents, turn on to full brightness nearly instantly, and use about an eighth the electricity of incandescents.

Re:At least it's not CFL

[Anonymous Coward]

He may not be a luddite, and actually have a good reason.

Quality of electricity varies. While you may be getting a near perfect sine wave at your point of consumption and have lighting which may exceed it's suggested lifespan, there are places where line noise is fairly bad. It might be a bad transformer or ground, the house might have a lot of dimmer light circuits, or somebody down the road with a high current draw doing something like arc welding. So yeah, under those scenarios CFLs might not last very long. In some places the electricity quality can be horrid enough to make radios near-useless, or computers unreliable (prone to BSODs and other errors) unless they get electricity from a power conditioned outlet. (Some places may have ok-ish electricity most of the time, but get brown-outs or surges alot. They'll still suffer similar problems.)

I wouldn't be too surprised that the guy who has no luck with CFLs also goes through computer power supplies at a higher rate than average, and has other electronics die much earlier (usually electrolytic capacitors and not just the ones with bad reputations) than normally expected. When you have that kind of problem, it's a clue you should call the electric company (or maybe your landlord) and complain.

Re:At least it's not CFL

[rubycodez]

I have some CFL that are four years old. I've also been the victim of crap ones that burned out in months, but haven't had that happen in last two years. I'd say give them another try buying a good respected grand, it's not bad like a few years ago.

Re:At least it's not CFL

[rubycodez]

haha, meant to type "brand" not grand. don't spend a kilobuck on lights like early LED adopters 8D

Re:At least it's not CFL

[Cammi]

What would be a good respected brand? I've hit every brand in Fred Meyers and Home Depot.

Re:At least it's not CFL

[rubycodez]

are you putting them in enclosed fixtures? that (heat buildup) or vibration (in ceiling fan, near hvac, etc.) will kill them. None of my Philips bulbs have burned out in two years

Re:At least it's not CFL

[SageMusings]

I looked at the lights you linked to. They have a 6K color temperature. That means a sinister, cold blueish light. I would like to replace a kitchen full of floodlights but I love the "warm" light I get from the incandescents. True, they do burn out more than I care for but they look great.

I've got some Philips LED lights with their "remote phosphor" tech in some standard lamps but I have not run across these in a floodlight form factor.

Re:This has been tried

[mdsolar]

These are street lights, not traffic lights.

Snow?

[asmkm22]

One of the side benefits of traditional bulbs is that the heat generated helps keep them clear of snow and ice. I don't think LED's generate enough. Anyone know how they are handling that?

Re:Snow?

[Anonymous Coward]

in my state, they realized this the hard way and switched everything back after a year.

i'm sure the unions of city employees paid to replace the bulbs are loving this. silly tax payers. pay more.

Re:Snow?

[Anonymous Coward]

Random fact: this is why lights run the other direction on railways. The red light is at the bottom where it can't be blocked by snow.

Re:Snow?

[necro81]

How often does the underside of a street lamp - ya know, the part that the light comes out from - get covered with snow?

Re:Snow?

[slew]

Random fact check:

Having the red on the bottom, instead of on the top as we are accustomed to with automobile traffic lights is a carryover from the old semaphore days, where the blade of an upper quadrant semaphore would "drop" from green to red as a failsafe/safety function in the event of a problem, equipment failure, or other mal-function.

http://www.railroadsignals.us/basics/basics4.htm#Format_

Re:This has been tried

[Amouth]

The is a big difference between Traffic lights and street lights.. the one they are talking about "street lights" are the ones facing downward towards the road, if there is enough snow to block the out, i doubt anyone will care at that point that they don't have a functioning street light.

Street Light Replacement

[Anonymous Coward]

I believe that Milwalkee. WI replaced all their streetlights with LED's some years ago. There is a problem with this in areas of the country that get snow. LED's generate no, or very little, heat. That means that the sun has to melt all the snow/ice from those streetlights after a storm, and that could take weeks or more. This causes a lot of confusion among drivers who are wondering who has the right of way. So, NYC should be careful, as the local constables might find themselves doing streetlight cleaning duties too.

Re:Street Light Replacement

[PPH]

Street Light != Traffic Signal

Yes, snow is a problem for traffic signals. Not so much for street lights which emit light downwards. And are not the same safety issue if they do get blocked.

Heaters can mitigate traffic signal blockage. And even with dozens of snowstorm days per year, will consume less energy than incandescents running 24 x 365.

Re:Street Light Replacement

[stewsters]

I live on the edge of Milwaukee. During the storms the whole thing can get covered, but usually the top part is visible because they have hoods.

The larger issue in total whiteout conditions is that you can't see far enough to see the light. The brighter the LED the higher chance you have to see it through the snow. The old bulbs just were not bright enough, even if they melted the snow faster.

If you are going to get in an accident because you cant read the light, then you need to slow down and spend some more time looking at other people's speed as they approach the intersection.

Re:Street Light Replacement

[jeffmflanagan]

Not only do they emit downward, a street light runs much hotter than a traffic light. I wouldn't be surprised if they need active cooling like LED plant lights do.

What about snow?

[Joe_Dragon]

it can block them and they don't have the heat to melt it.

Re:What about snow?

[EvilSS]

If it's snowing that hard, straight UP (these are street lights, not traffic signals) I think you'll have other, more dire issues to worry about.

Some of the first LED based signals not visible...

[drnb]

I recall a suburb of Los Angeles that installed some LEDs in traffic signals in the late 1980s. The was a visibility problem based on angle. At one light, maybe more but I only witnessed this once, when within about 10 feet of the limit line you could not see which light was illuminated. So the first car stopped at the light could not tell when it turned green.

One

[mdsolar]

But the light bulb has to be in a really solid state.

High Pressure Sodium are not Incandescent!

[Anonymous Coward]

They are gas discharge lamps, and are quite efficient. They do suck for light pollution, though. LED's with enough light output to replace them are also quite hot.

Re:This has been tried

[ebno-10db]

First, are you talking about traffic lights or street lights? Second, if the older style bulbs produced enough heat to de-ice the fixtures, then the combined power draw of the LED's and the heaters need be no more than the draw of the old lights. Even in tropical Minnesota, you don't get snow or ice ever day. They even have a month of something they call "summer".

Re:light pollution

[ebno-10db]

Ever hear of pedestrians?

Re:fp

[khallow]

There are two things to observe here. First, state and local governments have considerable latitude under the constitution. A number of things which would be unconstitutional for the federal government are allowed for them. Second, public lighting is one such allowed task.

Cue the Unintended Consequences

[TheRealHocusLocus]

Like 'dim' streetlights in Winter because they have frozen over, and (for cities that don't think of these things ahead of time) failure to install seasonally enabled heating units into the enclosure. All in all the human-time and effort of manufacturing and deploying these new solutions, along with the added heater circuit to make them useable, will really eat into that eco-energy difference equation.

Lots of eco initiatives these days come down to someone smiling and pointing to a little device that saves a few ergs of energy here and now, and just over the ridge there is a brand new factory making these things that is poisoning rivers and people with heavy metals. While very little energy is actually saved and unintended consequences pile up.

(Thank You Planet Saving Fluorescent Bulbs for saving the planet by seeding our landfills with elemental mercury instead of evil carbon. And giving me a HEADACHE whenever I am trapped in a room with you.)

And with LED light bulb revolution say goodbye to lots of radio communications. While the goofy things thrive on DC it is achieved through the use of really radio-noisy often insufficiently shielded switching power supplies and forced rectification. And brief high current pulses to 'cheat' higher light output without causing overheating.

Our city has LED traffic lights and even moderately strong FM stations disappear completely at intersections. I have no doubt that this interference affects emergency services' communications too, and that a whole lotta FCC Title 47, Part 15 violations are going on.

No one seems to care because people seem to be stupid when it comes to so-called eco-friendly product selling jobs. Sorry I so incorrigible about the subject, I do love the planet.

Re:Cue the Unintended Consequences

[mako1138]

Our city has LED traffic lights and even moderately strong FM stations disappear completely at intersections.

I very much doubt that LED lights produce output in the 90 - 110 MHz band. Switching supplies are generally 0.1-2 MHz.

But hey, call the FCC. They don't mess around when it comes to interfering devices.

Re:Cue the Unintended Consequences

[TheRealHocusLocus]

I very much doubt that LED lights produce output in the 90 - 110 MHz band. Switching supplies are generally 0.1-2 MHz. But hey, call the FCC. They don't mess around when it comes to interfering devices.

It's not your 'traditional' switching supply such as the ones that revolutionized electronics. The modular PC switching supply is inside a Faraday cage and has toroidal chokes so (almost) nothing but pure DC escapes the cage.

LED light systems tend to use HF pulsed current to drive the bulbs, and whether it's the bulb fixture or the pulsed power supply is actually in a separate control box with a long radiating (improperly shielded or grounded) wire connecting to the bulbs, I don't know. But the results are just as described here with noisy spectra shown here around 100Mhz.

Yeah, too bad I work for the city and there is a bread and butter conflict of interest.

Re:Cue the Unintended Consequences

[ChumpusRex2003]

I've had the same LED light interference problem. At work, at the exit to the car park, they have flashing warning lights. They recently switched from incandescent to LED. Now when you drive past, you lose FM reception when the lights are in the on phase.

Maybe it's a bad batch of lights, but the flashing isn't synchronised between all the bulbs, and you can hear different interference sounds on the radio and work out which bulb they relate to.

I've also had CFLs do the same thing, but those were super cheap junk ones.

That said, incandescents can also produce RF interference that in certain instances can be troublesome. I've certainly seen an incandescent bulb with a failing filament produce enough RF to render an MRI scanner unusable (this is actually a common fault found on MRI scanners with an "excessive image noise" service call). The mechanism is that a tiny develops at a weak point in the filament, but strikes an arc between filament ends. It is the arc (disturbed by vibrations in the filament, convection currents, etc.) that modulates the current in the MHz range.

Re:Cue the Unintended Consequences

[ebno-10db]

Thank You Planet Saving Fluorescent Bulbs for saving the planet by seeding our landfills with elemental mercury instead of evil carbon.

See p. 3 of http://www.energystar.gov/ia/partners/promotions/change_light/downloads/Fact_Sheet_Mercury.pdf

Coal burning power plants emit most of the mercury in this country. Using a CFL actually reduces the total mercury released into the environment because, in spite of containing a whopping 4mg of Hg, the electricity saved reduces Hg emissions by more than 4mg.

Re:Cue the Unintended Consequences

[TheRealHocusLocus]

Coal burning power plants emit most of the mercury in this country. Using a CFL actually reduces the total mercury released into the environment because, in spite of containing a whopping 4mg of Hg, the electricity saved reduces Hg emissions by more than 4mg.

I'm with you on the accuracy of those whole-planet-averaged figures in the link... but on the face of it they are comparing mercury dispersed by emissions over a wide area to a 4mg DOLLOP of mercury deposited in a landfill within the space of a few cubic inches. A substance with an EPA limit of 2 parts per billion in 'safe' drinking water.

That's like saying, we have found a way to offset 1000 statistically possible deaths of children that could (and on average, do happen) by placing 1000 land mines at playgrounds at random, per year. Comparing a poison slightly everywhere thing with a bunch of poison sure-things.

It doesn't end there. "Most mercury vapor inside fluorescent light bulbs becomes bound to the inside of the light bulb as it is used. EPA estimates that the rest of the mercury within a CFL -- about 11 percent -- is released into air or water when it is sent to a landfill, assuming the light bulb is broken." So 89% of this mercury is sitting there bound to glass. In my version of reality, 100% of the bulbs will be broken and reduced to small grain size. They will be buried and subject to extreme stresses, such as the plows of future generations in the mists of time. I also assume that everything in the landfill will be ingested by something or someone, some day. Topsoil everywhere is subject to this wormy process, and say in a thousand years, today's landfills could comprise mostly excellent topsoil. Except for the poor kid who ate that tomato with the EnergyStar logo on it.

I'm sort of leaning in the let our power plants more safely distribute mercury over a wide area while we read by incandescent light and come up with better ideas for power generation, kind of direction.

Re:Cue the Unintended Consequences

[Agripa]

I very much doubt that LED lights produce output in the 90 - 110 MHz band. Switching supplies are generally 0.1-2 MHz.

The switching power supplies typically used are not low noise resonant supplies so they have appreciable harmonic output. Being intended for consumer applications, they skimp on things like filtering and suppression.

But hey, call the FCC. They don't mess around when it comes to interfering devices.

The FCC has acted in some instances but even emissions that meet part 15 requirements can cause significant interference. The easiest solution is to switch to passive high voltage halogen incandescent bulbs.

Re:Cue the Unintended Consequences

[mako1138]

Well, I stand corrected. Thanks for the link.

you said "luminous vapor" huh huh

[Anonymous Coward]

"luminous vapor" is that some relative of "divine wind" ? snerk

Re:you said "luminous vapor" huh huh

[rubycodez]

and what does it mean if a southern woman says she has the "luminous vapors"?

Correction

[todfm]

White light inhibits night vision, not light toward the orange/red end of the spectrum.

Re:Some of the first LED based signals not visible

[Anonymous Coward]

My experience is the complete opposite. Overhere LED traffic signals have clear glass (the color comes from the light of the LEDs themselves) so you can easily see which lamp is lit, i.e., the one giving color. With traditional bulbs the cover glass is colored (the bulbs are white), so when the sun shines in the armature all lamps appear to be lit.

LEDs for traffic signals have many problems (being to damn bright at night, for example) but they are excellent for discerning what color the signal is currently projecting.

LEDs are often LESS efficient than Sodium lights!!

[Anonymous Coward]

My guess is the LED manufacturers have sold yet another city using fabricated cost "savings" numbers.
Notice how it fails to provide ANY information whatsoever on the total cost to replace all of those lights, and the estimated payback? It's pretty easy to guess why that is. It will cost much more than it will save, assumes LED reliability/lifespan numbers which are inflated, probably uses worst-case power efficiency numbers for the existing Sodium lights, and use the "directional" nature of LEDs to provide seemingly competitive lumen numbers when the lights actually produce noticeably less total light. Don't believe me? Google Luminous Efficacy (among other things).

For people replacing incandescent bulbs, LEDs are a huge reduction in power consumption, but sodium vapor lights are already efficient as hell, even rivaling LEDs in some cases.

LED light is weird

[arisvega]

Sodium light is a kind of light that the eye is very responsive to, and it is perfect for illuminating roads and freeways: any one of you that has done some nocturnal driving will agree that the difference between driving with your vehicle's headlights only on an otherwise pitch black road, with driving under sodium lights is huge.

Other technologies that produce bright light (Xenon, Halogen, all those 'energy saving' graded ones etc) are IMO are to bright to be of such use. Notice this: they are very bright, shine on things with wavelengths that make them look confusing and are in all good lights if you want to be seen, but not if you want to see. In the above example, imagine that you are driving headlights only, and then you enter a patch of that country road where the community thought would save money by replacing the sodiums with the bright white new energy saving ones, that watt-for-lumen are great. Then, all that happens is that the light blinds you, and since it is not really reflected back from the road that well not only you still cannot really see (like when you were driving in pitch black) but now you have a bright light on your face dazzling you

So, bright new tech lights: are bright? Yes. Make your vehicle look like some sort of inter-dimensional fearsome invader? Yes. Are effective? No. So please leave the sodium lights where they are and, if you do not hate other drivers, ban xenon lights from cars.

LEDs are good for trafic lights, as they are (kind of) directional and quite bright: however, if anyone has driven in underdeveloped countries where LEDs are now out of control and used a lot in advertising billboards (gasstations etc), will agree that it is very easy to confuse a, say, green or red light from said billboards with the real traffic light- so IMHO at least the red, green and orange leds should be regulated, especially on areas with traffic and traffic lights. Do not get me wrong, I am all into how they cyberpunk up the place, but bear in mind that you do not want some senior driver misinterpreting the local fish market's daily discounts shined on green LED letters for a green traffic light, and driving just ahead simply not registering the true traffic light and running you over (or worse). Not to mention problems in computer driving.

Finally, since there has been some progress in 'naturalizing' the 'white' LED color, it will be interesting to see if this progress is to the point where light reflect back from the street is in a manner that makes the surroundings look crisp, and aids the driver in being aware of them, which is the whole point.

Re:LED light is weird

[uncqual]

it is very easy to confuse a, say, green or red light from said billboards with the real traffic light- so IMHO at least the red, green and orange leds should be regulated, especially on areas with traffic and traffic lights.

I live in an area where yellow sodium roadway lights were installed many years ago (partially for "light pollution" reasons) and this created a problem much worse than any discreet billboard that I've seen.

Unfortunately, the roadway lights were almost exactly the same color as the yellow street traffic signal lights and lined up pretty well with the signal lights. When you were driving down some major streets at night you would see long lines of yellow lights (which were roadway lights) with an occasional red or green light (which were signal lights) mixed in. The problem was that when you saw a green light ahead and it turned yellow while you had glanced elsewhere, it virtually disappeared and required scanning among the street lights looking for the now camouflaged (yellow) signal light when you returned your gaze to the direct line of travel. After moving to the area, it took me a while to learn to make a mental note of exactly where upcoming green lights were so, if I glanced away, I could check in a fixed location to see if it had turned yellow. After a while it became second nature and I didn't even think about it anymore, but there were a few "oh shit, where did that red light come from" situations before that - esp. when traffic was very light so no driver ahead was putting on their brakes and providing a visual cue of a yellow signal light with their brake lights. I tend to glance around a lot when I drive anticipating possible upcoming issues (pedestrians, cross traffic that might be ignoring their signals, animals, bicyclists, kids, etc) so maybe I was more impacted by this problem than some.

Lots of people had this complaint but I guess it didn't cause enough serious accidents to warrant changing the roadway lights. Many or most of these roadway lights have been replaced with white LEDs. I now find it much less stressful to drive these roads at night and only now realize how distracting focusing on "remember the green light locations" effort was. I'm quite a bit more aware of pedestrians on sidewalks and bicyclists and the like now (which seems like a good thing). Although, the replacement LED roadway lighting seem to create more sharply defined "islands of light" with seemingly darker areas between the lights which make it a little harder to see objects off the side of the road in the dark patches - I suspect this is a combination of a narrower focus of the LEDs coupled with how the eyes adjust to the color and intensity of the sodium vs. LED lighting.

Is it me, or is this incredibly stupid?

[serviscope_minor]

I've heard about these "efficiency" measures before. People love to go all over LEDs because they're efficient (compared to incandascent bulbs). What they don't seem to realise is that no one in their right minds have used incandescents for bulk lighting in very many years.

All this will do is replace the reliable and efficient gas discharge lamps with less efficient LEDs.

But LEDs are fashionable so...

(yeah, the super efficient orange sodium ones actually need more lumens than the less efficient high pressure sodium ones once you factor human perception, but both types are substantially more efficient than LEDs)

New location for SNL

[mdsolar]

The sneering looks and jaundice pallor of the not-ready-for-prime-time-players in the opening credits will be completely sabotaged by the Mayor's plan. It's revenge for all those jokes at his expense I say.

How many people will die because of this?

[Anonymous Coward]

Illumination of Human spaces has now become a a political football, with dribbling sheeple THINKING they prove their 'green' credentials by backing solutions that endanger Human lives.

Every safety expert will tell you that the NUMBER ONE cause of whole classes of accidents is poor illumination. And the science of illumination is a complicated one, with apparent 'bright' bulbs having no useful brightness properties whatsoever, because they function in the wrong spectrum.

The move from filament bulbs was bad enough. A little power saving at the cost of new technology bulbs that contain the most toxic of materials, and have a functional fragility against complexity that means they often cost MORE than the filament bulbs they replace.

Here, in the UK, the government, in unannounced programs, flooded the supermarkets with the new compact fluorescent bulbs at prices as low as 15 cents (yes, less than one quarter of one dollar). This ensured people changed over with no complaint. Thus, we British got to discover the REAL life of these bulbs, and the picture isn't pretty. It is a miracle when one of these bulbs makes it past one year. Many fail within months. Sure, this is better than the cheapo filament bulbs we were using, by maybe a factor of three to five, but the BS about YEARS of like was a blatant lie.

And the new bulbs are of such varying quality. Some take ten seconds to reach acceptable brightness levels. Many have a poor colour spectrum. In fairness, the best come on quickly, and give pleasing light, but clearly no standards were set for the quality of the bulbs, and this should have happened BEFORE filament bulbs were effectively banned.

LEDs are worse, vastly worse. They are NOT bright. They get VERY hot if they are high powered. And the light they put out is of the very WORST kind, when the needs of the Human eye are taken into account. The Human eye HATES light composed from extremely narrow bandwidths. LEDs are useful precisely because they produce light of just ONE given frequency. A white LED, of course, is actually a composite of three of more LEDs, mixing very very narrow bandwidth signals like NOTHING that ever happens in nature. Our eyes simply did NOT evolve to work efficiently with such 'unusual' light sources.

Ever use one of the popular LED torches? Think about how the light feels. Yes, you can 'see', but you feel your eyes struggling to interpret the strange lighting effect.

As a person gets older, the optician tells them to read under BRIGHTER lighting. We see better in better lighting- and this idea is NOT as self-evident as it seems. A bright summer's day is insanely more bright than a fairly dim room in which we THINK we can see just as well. But our 'seeing' response time decreases rapidly as true brightness decreases, seriously affecting our reaction times, or simply our ability to 'notice' things.

LED street lighting, as that foul depravity Bloomberg well knows, will lead to significant increases in street accidents (and street crime). He doesn't care. What he does care about is that the braindead movers and shakers in NY will applaud him for this 'green' 'planet-saving' move. The dribbling hipsters who read the NYT will be pleased. And society will continue to roll backwards.

Here, in the UK, the equivalent Blair project is having ALL national garbage collection only once a fortnight (from the once a week service introduced by the Victorian Reformers who revolutionised public health in the UK). In regions where Blair has succeeded, illness associated with accumulation of garbage in people's homes and neighbourhoods, has rocketed. But Blair likewise sells his program on the back of 'green' thinking. George Soros tells you that filling your home/garden full of rotting garbage for two weeks is essential to save the planet, and most of you sheeple believe him. And when Team Soros/Blair/.Obama tell you that good night-time street illumination is abhorrent, and that you should embrace LEDs, flocks of sheeple come out to cheer.

Re:How many people will die because of this?

[roc97007]

Alright, I'm taking a chance here. Anyone who tells me "you shouldn't feed the trolls" is probably right.

> Here, in the UK, the government, in unannounced programs, flooded the supermarkets with the new compact fluorescent bulbs at prices as low as 15 cents (yes, less than one quarter of one dollar). This ensured people changed over with no complaint. Thus, we British got to discover the REAL life of these bulbs, and the picture isn't pretty. It is a miracle when one of these bulbs makes it past one year. Many fail within months. Sure, this is better than the cheapo filament bulbs we were using, by maybe a factor of three to five, but the BS about YEARS of like was a blatant lie.

I sense some exaggeration here. In the US CFLs never got that cheap in any size that I've ever seen, even at the mass discount stores. That must have been some massive government subsidy.

CFLs went through "value engineering" sometime in the late nineties, to the point where in the US the ones you buy six or eight at a time in blister packs don't last any longer than incandescent bulbs. Two of the three original CFLs I bought in the mid-nineties are still working. The blister-pack CFLs tend to last one to three years. (On average -- I've found that if they last the first three months, they're probably good for a couple years.)

The problem is not the technology, it's the implementation. The bulbs that ma and pa kettle are most likely to buy are the ones least likely to perform up to spec.

But, hey, wait a minute -- you're saying that CFLs that last one year only, are still better than the filament bulbs you were using? So how long do your filament bulbs last? What kind of crap lightbulb industry do you have going in the UK?

On the LED spectrum issue, that sounds plausible enough to be a Facebook forward, but not quite plausible enough to be true. Citation needed.

Re:How many people will die because of this?

[ChumpusRex2003]

I sense some exaggeration here. In the US CFLs never got that cheap in any size that I've ever seen, even at the mass discount stores. That must have been some massive government subsidy.

Not an exaggeration. Many stores would sell subsidised CFLs for about that price. Same with other energy saving products (I'd seen rolls of thermal insulation material - 10 yards, 6" thick on the shelves at hardware stores for about $2-3 each; but there were big warnings on the shelves which read something similar to the followion - warning! for personal domestic use only. Commercial use of this product is illegal. By purchasing this product, you certify that it will not be resold, used in the course of business or in the construction of a new building)

In fact, the energy suppliers had "energy reduction" targets to meet, and huge fines were levied if they didn't spend $x per year on assisting customers to use less energy. A common way for the energy companies to do this, was to buy massively cheap CFLs from China, claim the cost as a "green expense" and then just mail out unsolicited boxes full of CFLs to every customer. That really did happen, and the bulbs were the lowest possible grade available. The best bit, was that the energy companies could claim the cost of the CFLs as a "green expense", and the government would fund them. Where did the govt get the money from, it came from a surcharge on energy bills. It was even better for the energy company, if they could get a kick-back from the CFL vendor as part of a big order at list-price.

The cost of these "green projects" added to domestic energy bills comes to about $250 per household per year, accounting for about 15% of the total cost.

Re:How many people will die because of this?

[roc97007]

It's not just the monetary cost. Massive numbers of cheaply made CFLs in the hands of the populace practically guarantees that a significant number will help increase the mercury content in the local landfill. I know CFLs can be recycled, but who besides hippies actually do so? Most people will just tip them into the bin.

Re:How many people will die because of this?

[Anonymous Coward]

He's right about the spectrum of LEDs and it being displeasing.

Re:How many people will die because of this?

[roc97007]

He's right about the spectrum of LEDs and it being displeasing.

Hm. Wasn't aware of that. I'll have to do some more research on that. I never really liked the light from CFLs, to tell you the truth. It makes me wonder if the millionaires promoting CFLs are actually, you know, using them. Besides in the servant's quarters, that is.

Less efficient?

[Radical Moderate]

According to TFA, the city will save $6 million in energy costs per year. How does using less energy make them less efficient?

Induction lighting vs LED

[Anonymous Coward]

Will this stop the streetlights going off when 'psychic' people walk under them?

We prefer to be called "magnetic" :-)

And the answer is, thank god, yes it will stop that phenomenon. Existing streetlights use magnetic induction to induce light from the gas inside the bulb (thanks Tesla!) rather than internal filaments like incandescent bulbs. It's a lot less susceptible to wear failures than incandescents so bulbs last many times longer which is why it's so popular for streetlights.

LED streetlights do not use magnetic induction so those folks with truly "magnetic" personalities won't disrupt their lighting process.

Light Pollution

[supermachoman]

Although it may have some effect anyway, it would be great if NYC took the opportunity to reduce its light pollution at the same time. With lights that point down, they might be able to use even smaller LEDs, further saving money and decreasing time for ROI. The benefits to wildlife that are confused by the night lights and to citizens who might be able to see more of the wonder of the night sky are added bonuses.

I wish them luck

[roc97007]

Hopefully they'll work better than the LED stoplights that were put up around here a couple years ago. Everywhere you look you see LED arrays with 1/3 or more out or blinking madly.

I think LEDs are a very good idea, and I'm looking forward to the time it's cost effective at home to swap out these stupid CFLs, a technology that was (in my humble opinion) only ever a stopgap. But I wonder if current LED lamp technology is up to the extreme environments seen in outdoor applications? At least, after it's been value engineered and provided by the lowest bidder? Clearly, we're not getting stellar results with the current crop of stoplights.

Re:Some of the first LED based signals not visible

[SJHillman]

We have a few of these scattered around, but ours aren't LED. There's some kind of polarized filter that only makes the light visible within a certain angle. Annoying as all hell - fortunately, the ones I know of are all in real low-speed areas.

Special NYPD version

[Anonymous Coward]

Every one of them will have cameras and microphones.

Let me not

[mdsolar]

To the union of true memes
Admit impediment.Light is not light
Which alters when it alteration finds,
Or bends with the remover to remove:
O no! it is an ever-nixed dark
That looks on tempests and is never shaken;
It is the star to every wanders Central Park,
Whose worth's unknown, although his height be taken.
Light's not Time Square's fool, though rolling hips
Within his bending sickle's compass come:
Light alters not with his brief hours and blips,
But bears it out even to the edge of doom.
If this be error and upon me cited,
I never writ, nor no man ever lighted.

Re:Let me not

[wonkey_monkey]

Burma Shave.

LED synchronized peril sensitive shades

[An dochasac]

Incandescent? Are we stuck in a time warp? What city has had the money to waste on incandescent streetlights since the 1960s? LEDs are less efficient than the orange sodium streetlights but probably more efficient than the more common high pressure sodium. They have some key advantages. The first is that they can be physically much smaller than high voltage discharge lights which means it takes smaller optics to throw the light where you need it. But where light trespass remains, LEDs present an interesting option. I how many of us end up having to put ugly black-out shades in front of bedroom windows to keep unwanted streetlight glare from keeping us awake? What if an LCD shutter were synchronized to close exactly when street light LEDs were on and open when they're off. Suddenly you see the natural night sky from your bedroom window, are awakened by natural morning sunlight.

Re:LED synchronized peril sensitive shades

[EvilSS]

That would be fairly expensive and difficult to build (and you're assuming all of the LEDs visible to you, even in the same housing, would be on the exact same timing). That said, it would be seriously cool.

Re:LED synchronized peril sensitive shades

[geoskd]

Incandescent? Are we stuck in a time warp? What city has had the money to waste on incandescent streetlights since the 1960s? LEDs are less efficient than the orange sodium streetlights but probably more efficient than the more common high pressure sodium.

Any lamp that uses a ballast is going to be less efficient than an LED. The reason is that the ballast consumes 50% of the power at steady state. Even an electronic ballast is still going to draw around 25% of the power. LED power supplies by contrast run in the single digits.

Dark Sky approved? light pollution?

[Anonymous Coward]

A local municipality is also going to LED-based street lights, and they explicitly mention that what they're going with is:

LED street lights reduce the amount of light pollution. They are compliant with the International Dark-Sky Association (IDA) in terms of being dark-sky friendly by reducing sky glow.

http://www7.mississauga.ca/Documents/TW/ew/led-bg.pdf
http://www.mississauga.ca/portal/residents/ledstreetlighting

IDA-approved fixtures searchable at:

http://www.darksky.org/outdoorlighting

Anyone know what NYC is using?

Re:light pollution

[Anonymous Coward]

We call them targets.

no, actually they don't

[Chirs]

I live in the Canadian prairies, and I've seen traffic lights covered in snow because there wasn't enough heat to melt them. The city had to send out crews to clean the lights off.

It doesn't happen often, and the LED lights are still cheaper overall, but it does happen.

Re:no, actually they don't

[Joce640k]

Traffic lights have lots of small LEDs in them, they don't get very hot.

Big, powerful LEDs still generate plenty of heat. They're working on reducing it but it's still a problem. For a street light I suspect we're looking at 50-100W LEDs. That means plenty of snow melting ability.

Re:no, actually they don't

[plover]

Aeon Lighting Technology street lights are available in sizes up to 145 watts, and they emit 10,000 lumens. The entire top is a large aluminum finned heatsink. Indoors, in a 70 degree grow room*, with a little airflow, mine never get so warm that I can't grab them bare handed. The bottom glass panel in front of the chips stays room temperature. The tops would melt, but I bet they would form icicles readily.

*Before you call the cops, we raise orchids.

unfortunately....

[SuperBanana]

It only helps if somebody pushes at the correct time; but if the fixtures are being reevaluated in anything resembling a serious way, that's your best chance to get action on things like fixtures that point upward, ill-designed fixtures that don't target their output very well, and all the various other dubious lighting decisions that help add up to light pollution.

Unfortunately, the major impact around me is that our streets are now incredibly bright at night...they used the extra efficiency to make everything brighter, not use less power. God help you if you've got a bedroom that falls within the cone the new LED lights throw, too. My bedroom became lit like a supermarket, even with the shades down. It took four calls to the city before someone came out and re-tweaked the light.

Really, I wish people would pay attention to the studies that show that brighter != less crime/safer.

Re:unfortunately....

[Anonymous Coward]

I feel for you... I had a pizza place with the uberbright LED signage that I got sick of lighting up my bedroom so I placed foil in the windows. The landlord thought we were running a grow-op.

Re:fp

[AK Marc]

Have you never heard of the 10th Amendment?

Piggyback Wifi

[Phoenix666]

Pity they don't piggyback public wifi onto the lights with those LED wifi bulbs in the news a couple days back. It nicely resolves right-of-way concerns and power supply requirements.

Aren't HPS lights more efficient than LED?

[kawabago]

I was under the impression that LED's were about 93% efficient at converting energy to light and that High Pressure Sodium bulbs used in street lights were about 95% efficient.

Re:Aren't HPS lights more efficient than LED?

[AnotherBlackHat]

High Pressure Sodium street lights are around 100 lumens per watt. HPS can be as low as 50 or as high as 150 depending various factors.
LED street lights are also around 100 lumens per watt, but they vary even more widely.
A theoretically perfect light source with a wavelength of 555 nm would emit 683 lumens per watt, so both LED and HPS are about 15% efficient.

Even the best HPS is less than 25% efficient, and the best LED isn't much better at just over 35%.
At 14.5 lumens per watt, an A19 incandescent 60 Watt bulb (i.e. a "normal" bulb) comes the closest to 97%, but it's 97% inefficient.

Re:light pollution

[Anonymous Coward]

Ever hear of pedestrians?

What does religion have to do with it?

You're kidding, right?

[Overzeetop]

I presume you are utterly colorblind.

The reason your headlights don't look like street lights is the illumination is entirely different - narrow cone illumination in front of you (nor peripheral coverage) vs. overhead flood lighting. If you'd ever been in a warehouse with LP Sodium lights you'd realize what a horrid source of lighting it is. The only reason to use Sodium is the trade off of fantastic operational cost vs crappy light and astronomy-killing wavelength pollution.

If their numbers are correct, they will spend $10M to save 700,000/yr (14M over 20 years) - that's an annualized rate of return of 3.45%. It's good that it's positive, but it's a pretty slim actual savings.

Why

[Anonymous Coward]

Metal halide and LPS bulbs are still more efficient than LED. Why would the install fixtures with reduced efficiency?

How much will it cost?

[edibobb]

How much will it cost to make the change? This should be part of the "No Brainer" decision. I agree that the change is good, but rational people would expect to know the time it takes to recover the initial investment.

I wonder if thieves will steal these

[peter303]

They are at least an order of magnitude more costly than incandescents (About 20x at Home depot this month). Theives steal anything they can take and resell.

LED street lights are over 100 lumens / watt.

[AnotherBlackHat]

Yes, it's you - you are so last year.

We aren't talking about those cheap $10 bulbs from Home Depot that get a measly 84 lumens/watt.

Commercially available LED street lights like Cree's LEDway have been over 100 lumens / watt since 2013-03, making them more efficient than high pressure sodium even without considering their arguably superior focusing and CRI.

Re:LED street lights are over 100 lumens / watt.

[serviscope_minor]

have been over 100 lumens / watt since 2013-03, making them more efficient than high pressure sodium

You mean more efficient than the best high pressure sodium lamps at 150 lumens / watt or the best low pressure ones at 200? Even good, white metal halide lamps can exceed 100 lpw.

Sodium lamps are currently the most efficient lighting source. I expect LEDs will exceed them some day, but that day isn't yet.

I'm going to believe the easily verifiable figures for efficiency over unverified numbers from a politician trying to make himself look good.

Incandescent Sodium

[Peter Simpson]

Street lights:
LED (should) last longer than Sodium for about the same number of lumens per watt
Sodium draws less power than Mercury vapor
Mercury draws lots less power than incandescent (which no one uses any more)

Traffic lights:
A good quality (many of the early ones weren't) LED signal will last far longer than an incandescent bulb
I've never seen the "snow obscures signal and doesn't melt" problem here in Boston, but it seems plausible

R.I.P

[Anonymous Coward]

R.I.P Mr. Thomas Edison!

You did have a bright idea that lasted for many decades! Now, I have an LED bulb over my head!!

Got `em.

[skelly33]

The LED street lights have shown up in my neighborhood in CA over the last year. Frankly, now that I see them in place, I don't like them one bit, for the same reason that I don't like LED brake lights on vehicles (as if we need to conserve energy on brake lights, please!): each super-bright LED is a very intense point of light which immediately makes its mark on my retinas and I see the spots for quite a while. I can't be the only one with this problem, and I can't imagine it not having a long-term effect. Sodium, fluorescent, and other kinds of lights seem to more often be accompanied with some sort of diffusion that eliminates the high intensity pinpoints from direct view. Not the street lights: one glance at those and I get a lovely 8x20 matrix of dots in my field of view for the next several minutes (or a 1x40 string in the case of brake lights). I think some improvements need to be made before they continue rolling out en masse.

Tangentially related, I don't particularly feel like we need street lights on all night long. What if we just lit up side walks with low posts (perhaps lower even than the FOV of a typical driver - enough to light the path and cast enough ambient light for pedestrians to take advantage of, but WAY lower power than the street lights, and with no intent of lighting the entire community? If my car's headlights are sufficient in the back woods where there are no street lights to drive safely on the most treacherous of roads, then why would I need street lights to guide my way in town where the roads are all flat and predictable? I, for one, would welcome a far less lit night sky for star-gazing and total overall reduction in energy consumption.

Re:This has been tried

[Valdrax]

They even have a month of something they call "summer".

I thought the seasons were Cold & Wet, Cold & Dry, Cold & Wet 2: Electric Boogaloo, and Bugs.

Los Angeles has them, and they're fine

[Anonymous Coward]

Los Angeles has replaced nearly all its streetlights with LEDs, and they work fine.

See http://bsl.lacity.org/

Probably none.

[Valdrax]

Every safety expert will tell you that the NUMBER ONE cause of whole classes of accidents is poor illumination.

No they won't. According to the US's NHTSA, the top cause of accidents is inattention. Nearly 41% of accidents involved distraction within 3 seconds of the accident. 71% of accidents occur during the day. Night driving in dark, unlit conditions accounts for only 12% of accidents.

So, unless "whole classes of accidents" means "just the ones with poor illumination," your statement is unlikely to be meaningfully true.

LEDs are worse, vastly worse. They are NOT bright. They get VERY hot if they are high powered. And the light they put out is of the very WORST kind, when the needs of the Human eye are taken into account. The Human eye HATES light composed from extremely narrow bandwidths.

Citation needed.

And when Team Soros/Blair/.Obama tell you...

Yeah, yeah. And tell us what the lizard jews running Zurich want the Illuminati to do next.

Reliability

[justthinkit]

Reliability is a word not yet appearing in these comments. I'm thinking back to when the first LED street lights were installed, and there was a proliferation of lights where one portion of the light had failed, creating a very visible and ugly effect. I don't see this today and I'm sure it is because the initial batches of POS lights have been uninstalled.

If this is going to happen again, to lights costing a $1,000+ apiece, times 250,000...we have a potential waste of $250M. Hopefully my city will wait for the good LEDs to make their way to market.

Re:light pollution

[wonkey_monkey]

pedestrians?

They won't be able to molest kids if they can't find them in the dark.

Seattle as a model failure for LED Streetlights

[Anonymous Coward]

They've done a lot of replacements in Seattle neighborhoods under the auspices of energy efficiency as well. Of course, they've also upped the lumen output and changed the color of the night. For a savings of about 1 cent per household per night I've encountered many people who have spent hundreds of dollars on blackout shades after having trouble sleeping because of all the excess light, and blue light at that. Apparently the LED companies don't "get it" that human wakefulness depends on the wavelengths of light that stream to our retinas while we sleep and tone-down the blue portions of the spectrum...

We can only hope that NYC makes better choices than Seattle has.

Was LED Droop solved?

[Anonymous Coward]

If the output of a LED bulb droops by 20% in a few weeks, then most/all the saving evaporate rather quickly:
http://news.rpi.edu/content/2013/07/30/researchers-identify-cause-led-%E2%80%9Cefficiency-droop%E2%80%9D

Better late than never, I guess!

[Anonymous Coward]

Even most part of the US have done this already. In countries that dominate world LED production (Taiwan, Singapore, China, ....) all traffic lights have been LED for well over 10 years. Hell they put LEDs in the sidewalk concrete and the sides of buildings.

Shields

[brunnegd]

I hope the new fixtures have shields to direct light down, keep it out of the sky. An even more efficient approach to lighting.

Really? Incandescent?

[Mirar]

If the city really still uses incandescent light bulbs - someone confirm? - then indeed savings are ahead. Most of the rest of the world switched to high-pressure sodium years ago.

If the city already switched to sodium, it's hard to see that it's an improvement:

low-pressure sodium: up to 183 lm/W [http://www.sla.net.au/sites/default/files/SLP.Pdf ]

high-pressure sodium vapour lamp: 93 lm/W [http://www.unep.org/climatechange/mitigation/sean-cc/Portals/141/doc_resources/TrainingEEtechnologies/EE%20Lighting_Asthana.pdf]

LED: 100-120 lm/W according to manufacturers

or worse: 50 lm/W
[http://www.ledlightingexplained.com/led-lighting-myths/]

LED driver failures

[mrflash818]

If you see an LED stoplight where a chunk of the LEDs seem out, or are blinking wildly, it is likely the circuit that supplies electricity to those LEDs, the LED driver circuit, is what is actually failing, not the LEDs themselves.

Ideally they can be swapped out and the light returned to service, but certainly does lower the hopes that cities had in installing them: to reduce replacement and maintenance costs.

Re:LED driver failures

[roc97007]

If you see an LED stoplight where a chunk of the LEDs seem out, or are blinking wildly, it is likely the circuit that supplies electricity to those LEDs, the LED driver circuit, is what is actually failing, not the LEDs themselves.

Ideally they can be swapped out and the light returned to service, but certainly does lower the hopes that cities had in installing them: to reduce replacement and maintenance costs.

Ok, so, I haven't seen the design... is there a driver for each individual LED? I ask because the failures I've documented with a camera (on my blog, which I am reluctant to share here lest it be slashdotted) are seemingly random, ragged patterns.

I agree, it's probably not the LED itself. I was leaning towards TCE expansion/contraction causing cracks in the circuit board.

But for whatever reason, they don't seem to be very robust. Which is a little odd as the technology should be practically bullet proof if designed correctly.

Streetlights

[Anonymous Coward]

NYC just changed all of their streetlights in 2010!! The city had a random combination of 150 to 400 watt HPS fixtures, including some mercury vapors. Even the damn streetlights in NYC changed! That was one thing I admired about NYC, the different types of streetlight fixtures. They now are all 100-150 watt General Electric luminaries. SMDH