Detroit’s LED streetlights going dark after a few years
detroitnews.com
detroitnews.com
We urgently need some standards over electronics longevity, like the ones we have over power consumption.
I've gone in and returned probably a dozen such bulbs without any receipt, just saying "hey, these are your bulbs, you say 10 year warranty, this is broken and they only came to the market 5 years ago".
Of course this only works now in the early days. If the date of manufacture was printed along with the store name, it'd be problem solved forever.
> EU law also stipulates that you must give the consumer a minimum 2-year guarantee (legal guarantee) as a protection against faulty goods, or goods that don't look or work as advertised. In some countries national law may require you to provide longer guarantees.
If the package says something like "10,000 hours" and you haven't run it for that many hours and it dies in 2 years you can get it replaced by EU law
Yes, and that mean an incandescent bulb is 100% efficient: all power used is converted to light+heat. But other systems can be more than 100% efficient - they can add more heat than the power they use. Or, in the case of central heating from waste, unused heat will probably have to be pumped off anyway, so using it is energetically free.
Yes.
> What is the other source of inefficiency that makes them inefficient for this by-use?
Heat pumps are more efficient at heating than resistive elements because they take advantage of already existing environmental heat, whereas resistive heating has to generate all of the heat directly.
To increase efficiency, many cities run cogeneration plant, where the waste heat from electricity generation is piped into municipal heating system. In this scenario, the incandescent bulbs are wasteful way to heat the house, because it only uses 20-40% of the energy of the fuel, while municipal heating uses only waste energy to produce eletricity.
Hmm.
edit - and small, you could dot a lot of spare leds around without a problem.
He does tear downs of consumer grade led lights for (mostly for the European/UK market since that's where he's based). It's interesting seeing some the iterations that "cheap" devices go through as they are refined to the bare minimum of cost.
Edit: Sometimes LED current is actually high-current AC with a square waveform, especially with dimmable bulbs. This is fine provided it's well-regulated and protected from spikes.
LED tolerance doesn't really play into the scenario here. It's all about the driver.
As a result you generally need to wire them in series rather than in parallel so that the current can’t differentially select the LED of least resistance and cause thermal runaway. Series wiring in turn cranks up the voltage requirements necessary to drive lots of power, requiring more costly components.
Oh, forgot to mention, different color LEDs have different forward current requirements, so have fun building a good and cheap driver for 100W RGBW.
LEDs can easily be wired in parallel as long as you use resistors to balance them. If you dial in the voltage real tight you can get away with minimal losses in the resistors. If you are really fancy you can use transistors and shunts to actively balance then for even more efficiency.
The reason why series is more popular is actually because it is cheaper and(can be) more efficient. It's not uncommon for cheap led lights to just throw 40 leds in series with a dead simple "driver" (read: rectifier, smoothing caps, and a resistor)to run it right off 120V.
Similar issues can affect parallel. If one LED fails open and you don’t adjust the current, you’ll start overdriving the other LEDs and shorten their lives...potentially dramatically. You could add current sense and shut down the board or even adjust current, but then costs go up.
Typical LED strips are small serial strings of typically three lights connected in parallel. That way if something goes wrong you knly lose a few lights.
Good caps are expensive, and cheap alternatives are plentiful. Nine out of ten off the shelf LED drivers have cheap ass chinese caps in them.
You also need good overall thermal design, which means good heatsinks, which means chunks of metal, which of course are expensive.
It's entirely possible to make a driver that will outlive any LED it powers, it's just that nobody wants to pay for it.
I dunno about the shoes but Snap On isn't really great quality anymore. The Chinese have closed the gap. Fluke, Honda and to a greater extent Toyota are all steadily moving toward more disposable products.
The cheap Chinese kitchen stuff you mentioned is actually the most durable item because in order to stand up to occasional abuse without being damaged it needs to be so overbuild that it can last lifetimes. That's simply a function of the design requirements, not any particular manufacturer's attitude though.
You raise a good point though. Its very hard to evaluate a product that's supposed to last a life time, until you've had it a lifetime. So the benefits to a manufacturer are slow to accrue, but they're also slow to dispel, so there's massive temptation to cut corners and rely on that reputation, while lowering costs.
It'll be interesting to see how the EV market works out. You have a product that could last for life (with battery replacements). I suspect they'll be viewed, by the manufacturer at least, as disposable when the battery goes.
Not sure what availability is like in the US and other countries though.
Really? As an anecdatum, I replaced most of the lights in our house with cheap LED globes a few years ago and so far not a single one has failed. Maybe I'm just lucky. CCFL lamps, however, have been terrible while using 3x the power.
The thing is, as bad as the waste is — I'd much rather chuck out LED bulbs than CFL/FL.
Or did it. I noticed that if you keep it on long enough after it turns off it will come back on. I think the fuse or something gets heated and it stops until it cools.
My theory is that they test them in lamps upright. I’ve noticed much higher failure rates with fixture based and inverted bulbs.
The streetlights are putting out a lot more power (and thus heat), and thus need much better thermal management, which it appears they don't actually have.
I've had 2 led failures since I went fully led 8 years ago, one went soon after purchase and looks like a manufacturing defect, so that leaves one that could potentially be heat. Most of them are no names, most recent are Phillips though.
Heat is really ungood for LED's and the drive circuits[1]. With a recessed fixture there is no place for the heat to escape.
[1] I've noticed excepting electrolytic caps electronics that doesn't get warm, basically lasts forever.
Nothing happened to incandescent bulbs or fan.
Luckily I had a power strip with surge protection on which all my expensive equipment were hooked. Power strip blew up but saved my equipments.
Edit: if you open your mobile charger you'll see 400v capacitor right after the bridge rectifier. On 220v mains this circuit needs to handle 310v but can survive upto 400v peak to peak.
But power surges can be more than 400v, and similarly the circuit in CFL or LED driver can only survive very low band of power surges.
Unless you've surge protectors in your house actively scrubbing mains dirty electricity, you'll see your LEDs will not survive long enough.
LEDs don't survive because driver quality sucks and can't deal with mains fluctuation, and LEDs are very sensitive to temperature and voltage, little change in temperate or voltage, will drastically change the current going through the LED and will burn them.
I solved the problem of continuously burning my LEDs by running low voltage DC lines and using 3 level surge protection in my house. Now, I've several mounts for LEDs with fans which keep the temperature in control.
You need serious cooling for anything over 5w led. I am using aluminium heat sink with fan and my LEDs don't die.
This is key and how all houses should be designed. All lighting circuits should be low voltage DC. This makes the individual drivers much simpler.
But reality was that you had the choice between dozens and dozens of brands you never heard of before, and chances were high the one you picked just sucked. Amazon reviews were worthless because most people wrote them weeks after buying, very few people would come back if they failed after a year or two, and often by that time the company might have rebranded everything so you wouldn't know anyways. Not to mention how many people apparently are unable to recognize that atrocious 100hz flicker if the company cheaped out completely.
Long story short, all my led bulbs since 2015 are osram or Philips and I haven't been disappointed since.
Why are all these cities undertaking huge renovations to go to LEDs?
If I use conservative estimates (10 hrs per day, "few years" = 4), then these LEDs are failing at under 15,000 hrs.
A 10 year warranty probably means the average is longer than 10 years.
And the point of specific warm LEDs is avoiding higher frequencies. It's not the same as mixing light from red and green LEDs.
> The sodium lamps use roughly the same amount of energy
"The new LED lights will consume 50% less electricity than the existing High Pressure Sodium (HPS) lights. When fully implemented the new fixtures are estimated to save the City $10 million a year in utility costs. By consuming less electricity, the City is helping reduce its carbon footprint. LED fixtures last up to two-three times longer than HPS fixtures."
> cast a light quality that is better for both humans and animals
"Compared to HPS lights, LED lights make it easier to see the contrast between an object and its surroundings; so one can quickly and more accurately identify people, vehicles, road debris or other things on the street or sidewalk. The whiter light also enhances peripheral vision and improves the quality of video resolution"
[1] http://chicagosmartlighting-chicago.opendata.arcgis.com/page...
Not sure how they're going to achieve that. LEDs produce 37 to 120 lumens/watt, HPS and LPS produce 50 to 160 lumens/watt. The sodium lamps lose out only because they are omnidirectional and a mirror is required to aim their light down, so some of the light is blocked by the lamp itself. But overall LEDs don't save 50%.
> Compared to HPS lights, LED lights make it easier to see the contrast between an object and its surroundings;
This is true, but it's a red herring. While that might increase safety in one regard, the light cast by an LED is very blue, which affects both humans and animals, making their bodies think it is still daytime. This disturbs the natural hunting and migration patterns of the animals, and more importantly, disturbs the sleep of the humans.
Sleep deprived humans are for more dangerous than ones that have a harder time detecting object edges.
Also, shouldn't we consider quality of life issues when making these changes, and not just efficiency?
For blue LEDs, sure.
LEDs today can be made to emit any color all the way up to pure white.
In practice, all “white” LEDs (even the very warm ones) consist of a blue LED + a yellow phosphor, and for efficiency reasons in contexts like street lamps the phosphor is never set up to absorb all of the blue light, which means that the intense blue glare inevitably is there clobbering people’s night vision.
It would be possible to make LEDs appropriate for street lighting, but the lighting industry does not currently sell them.
They're getting cheap enough now (insanely cheap, in fact) that it is now feasible to fit street lamps with them.
Imagine the use that could be had if you could change the color of street lamps on demand. Guide vehicles in an evacuation, direct police to a particular location by coloring a lamp red, or something. Direct ambulance by coloring a lamp blue. I don't know. Seems like it could be useful.
It's less about the LED being suitable for street lamps than it is about the lensing chosen to direct the light to the proper locations under the lamp.
This is better for displays (or maybe stage lighting, art projects, fancy effects, ...) but not better for standard room lighting, and definitely not better for street lamps.
Using street lamps as information sharing devices makes a cute 1 minute demo, and sounds cool to sci-fi fans, but in practice would be gimmicky and not very effective.
The actual light coming out of “RGB” LEDs necessarily has a quite spiky spectrum, because they overemphasize supporting a wide color gamut. They also include more short wavelength light than is desirable for nighttime use (compromising night vision). e.g. https://cdn2.goughlui.com/wp-content/uploads/2015/12/hue-def...
> It's less about the LED being suitable for street lamps than it is about the lensing chosen to direct the light to the proper locations under the lamp.
I’m not quite sure what you are trying to say with this bit. For best lighting, there would be more lamps spaced closer together, placed at lower height, orange in color and individually dim, and diffused and well shielded from the side. These are all features which would be helpful for any type of lamp.
Unfortunately, the LED streetlamps adopted by cities are high up, widely spaced, much too blue, extremely intense, and not shielded from the side.
Less lumens. The lighting industry has successfully scammed the public into buying less efficient LED lighting with greenwashing.
The availability of cheap overkill LEDs is making cities worse... It's especially noticeable when walking at night in areas with restaurants and nightlife.
It's not just street lightning, but also private buildings (maybe they're after that prison yard look?) and cars (giving pedestrians and drivers free lasik, nice!).
Maybe these are OK for highways, but this sucks big time for everywhere else.
(For the same reason, LED car headlamps are horrifically bad for everyone other driver on the road.)
They don't realize there's a problem, or they've just spent a lot of money "upgrading" their vehicle fleets and don't want to consider that they've been ripped off.
The vehicles of my local police departments and state highway patrol have had all their safe incandescents replaced with blinding LEDs. Flashing blue LEDs on the roof of the cars, flashing blue-white LEDs mixed in with the flashing red and blue LEDs, and blinding amber LEDs to warn of road hazards. The vehicles have blue-white spot lights, and sometimes they have blue-white light bars on to blind everyone within a quarter mile of the scene.
I think the police officers think their new lights are cool.
OSHA [0] should intervene, but I don't know that I can make a complaint on the police officers' behalf...
[0] https://en.wikipedia.org/wiki/Occupational_Safety_and_Health...
I did a quick search and found their RFI [1]. They are reducing electricity consumption by 50% because they are replacing 40+ year old light fixtures that have 70+ year old wiring:
> In the 1950s, the City installed new lighting infrastructure citywide; utilizing mercury vapor lamps. In the 1970s, the City performed a system-wide lighting upgrade; replacing the mercury vapor fixtures with high-pressure sodium fixtures. This project was only a lamp head upgrade; the City retained the 1950s era electrical poles, wiring, and infrastructure. Over the years as underground wiring has failed, City crews have often replaced underground wiring with aerial wiring. In 2009, using available federal funding, the City replaced high-pressure sodium streetlight fixtures with more energy efficient ceramic metal halide fixtures along Western Avenue and Lake Shore Drive. That fixture replacement undertaking did not include upgrading the electrical infrastructure.
[1] http://chicagoinfrastructure.org/wp-content/uploads/2014/10/...
A new supermarket by me installed all LEDs for the parking lot, and used warm white ones. You wouldn't even be able to tell they aren't incandescent without looking. I really hope more word gets out to not use "bright white" LEDs for outdoor lighting.
Street lighting does not reduce car-car collisions, and in most conditions, does not reduce collisions at all. The major exceptions were pedestrian crossings without traffic lights, and when illuminating curves or obstacles. I don't recall finding a comparison between overhead street lighting and reflectors or marker lights for curves and obstacles, but it seems to me those might provide the same benefit with less glare, energy consumption, and light pollution.
I want to be woken up by the light of the rising sun. Curtains that keep up street lighting also keep out the sun.
IMHO there's far too much street illumination anyway.
That doesn't sound right, source?
edit: https://www.stouchlighting.com/blog/led-vs-hps-lps-high-and-...
Why would LEDs put Sodium vapor lights out of business? Sodium lamps have the worst color rendering of any bulb. They produce a dark yellow glow which is generally a very low quality light. Additionally, there are serious waste disposal issues with sodium lamps. In particular, they have been known to start fires in the event that the lamp is broken and the sodium metal is exposed. The sodium can catch fire even in the event that the lamp is broken on the ground. For this reason it is safest to break sodium lights under water and then to subsequently dispose of the destroyed bulb. Lastly, HPS and LPS lights are monochromatic, so they can mess with your color vision if you look at them for an extended period of time.
Perhaps more importantly, in the last few years LED efficiency has surpassed that of even LPS and HPS lights and its efficiency improvements are progressing at a much more rapid rate. The largest selling point of LPS and HPS lights is the cheap selling price, the high energy efficiency (low operating costs), and the relatively long lifespan. LPS and HPS still retain these advantages over most conventional bulbs but they lose on all three counts to LEDs. In some areas (e.g. lifespan) they are drastically inferior to LEDs. The extremely low maintenance and replacement costs with LEDs is actually a major cost benefit over the long term. LED lifespan can be greater than 100,000 hours (more than four times that of LPS or HPS). Having to purchase one bulb versus 3 or 4 bulbs over the course of time is a significant selling point for LEDs. The bottom line is that having lost their traditional advantage of being the most energy efficient bulb on the market, there’s very little reason to use a sodium vapor light when LED lighting is available.
The very narrow spectrum benefits astronomers because it's easy to filter out light pollution from low-pressure sodium lamps.
I assume animals have evolved to deal with that.
Relevant Technology Connections video: https://www.youtube.com/watch?v=wIC-iGDTU40
Street lamps should include as little short wavelength light as possible and should otherwise be as even (this means: use more lamps spaced closer and placed at lower height, diffused and shielded from the side) and dim as possible, so as to avoid causing distracting glare in people’s peripheral vision, and avoid causing high contrast between areas directly under the lamps and areas in shadow. Human vision is amazingly good at adapting to very low light levels. After adaptation, humans can navigate the environment by e.g. starlight, but at any rate can see just fine under sodium lamps.
But the way cities roll out LED street lighting is to put a small number of widely spaced very intense blue lamps high up in the air, not diffused and with little shielding from the side.
Every parameter has been optimized (pessimized?) for blasting away people’s night vision and causing enough glare to make seeing into the shadows all but impossible.
Extremely low color temperature LEDs would be fine, but don’t really have any especially great advantage over sodium lamps. (Except maybe for getting some government subsidy money?)
For example, while sure, OSRAM advertises/sells street lighting LEDs that are broad spectrum (131 lm/W) [1] they also sell single-color LEDs like these amber ones that are actually even more efficient (141 lm/W) with an emission range almost entirely within 600-640nm [2].
This is almost as tight a range as LPS (which looks like about 580-610nm) [3], and it beats out HPS considerably [4].
[1] https://www.mouser.com/new/osram/osram-street-lighting-solut...
[2] https://www.osram.com/os/ecat/OSLON%C2%AE%20SSL%20150%20GA%2...
[3] http://www.flagstaffdarkskies.org/low-pressure-sodium-lighti...
[4] https://en.wikipedia.org/wiki/Sodium-vapor_lamp#/media/File:...
This sounds like dangerous advice. Sodium reacts exothermically with water producing NaOH. In practice this means it will violently boil and splatter strongly alkaline lye around. (Source: this is how my high school chem prof lost some of his eyesight.)
Short answer: it's complicated.
From what I understand one big driver is federal subsidy money. I’m not an expert on the financing though.
Also a ton of industry marketing / propaganda. And poorly designed recommendations / regulations for road engineering.
Nearly every brand I've tried started flickering or dimming or failing in some way after 2 years of use despite the box touting 10+ years of life. Disappointing to say the least.
Also, the IKEA ones seem fine.
Are you sure your brands are original? Be careful if you buy in places like Amazon, at prices too good to be true, because they will be Chinese clones.
They're always underspecced so you get a little square of light directly under the street lamp, but that's about it. Looks like early 3D videogame lighting.
I was recently in LA and drove down to long beach, every other streetlight was dead.
Maybe by the time they figure out what a disaster this is and have to redo it all, they'll do a better job at cutting out the light pollution.
Ugh. I hope the right long-term changes happen where they are needed.
I noticed the article noted that units supplied from Cree, used in other areas of the country, are performing up to standards and not dying like the ones in Detroit from Leotek.
To me, it's one thing if you're a hobbyist or something looking to save some bucks so you buy cheap LEDs drop shipped from China via Ali Express, from some rando vendor and manufacturer. If one or more die, usually it's not a big deal.
But if you're building a product, or installing something commercially or at a government level, you should do yourself a favor and not go for the lowest cost. While I'm sure Cree LEDs are probably manufactured in China, too, they are likely made to a much different standard.
For all we know, Leotek is using relabeled parts that are marginal or failed to pass QA inspections at other manufacturers, but they got a great deal on barrels of rejected LEDs, and decided to go into the commercial light manufacturing business.
I honestly don't know who or what the company is. What I do know is the name "Cree" and their quality (which also comes with a hefty $$$ price/cost) - but "Leotek" sounds like a company you'd find in a random search on Alibaba.
Maybe that's an unfair assessment; sometimes when ordering a large volume of parts, you'll get entire batches that fail in the real world when they passed QA fine...
[1] https://en.wikipedia.org/wiki/San_Francisco–Oakland_Bay_Brid...
The vendor promised a certain warranty. If the due diligence was done, and the cities were happy with the price paid and the terms of the warranty, hopefully the vendor stands behind the product.
But: '"This deadline has come and gone, and Leotek ceased all communications with the PLA," the suit reads.'
Company website has both: "Leotek Electronics USA LLC", and "Leotek is a subsidiary of LiteOn Technology Corp."
https://leotek.com/about-leotek/
IANAL, but that would seem to be a contradiction.
I'd give generous odds on a bet that the end of Leotek, and Detroit's recovery efforts, are not far in the future.
It appears the LLC was formed in 2014, after Leotek merged with LiteOn: https://www.liteon.com/storage/document/document/engfile-d21...
They’re being replaced under warranty.
Am I wrong to think that not all is lost, all they need to do is change the bulb...and the major investments were the poles? Not trivial of course, but also not a total $180m loss
A union journeyman costs ~$65/hr in Detroit
A 50/50 split between labor and materials isn't unusual for electrical work
They bought from multiple vendors, following procurement rules, from a company in San Jose, CA (and a division of Lite-On), from a company supplying lights across the country, and the design was flawed, not the parts. Don't blame them when you know nothing of the situation.
> Founded in 1992 in California's Silicon Valley, Leotek is a leading global manufacturer of street, roadway, area and signal lighting products and solutions. Leotek streetlights were first designed and delivered in 2007 and the company now has more than 1.5 million installations across North America.
> In 2013 Leotek established a manufacturing facility in San Jose, California. [...] Leotek's lighting products are subject to 100% 2-hour "burn-in" reliability testing and comprehensive performance testing prior to shipment.
>中国·山东山东佳虹电子科技有限公司是由台湾1992年建厂的母公司—光林电子股份有限公司(LEOTEK ELECTRONICS CORPORATION)，经毛里求斯投资设立，专业从事LED应用产品研发及制造，属高科技型并拥有进出口权的生产企业。 #Shandong Jiahong Electronics.
>研制、开发和生产新型电子元器件、电力电子元器件、光电器件、新型显示屏器件、发光二极管（LED）应用产品、LED交通信号产品、交通设备及相关电子产品，并销售公司上述所列自产产品。 #R&D and production.
Oh that's a popular trick now, you'd think China would get around its bad reputation.
https://arstechnica.com/tech-policy/2017/03/lawsuit-3-5m-ind...
Like this California startup Acumen Robot Intelligence is actually a Shenzhen company subsidized by its local government.
(The Chinese doesn't particularly help - I can't read much of it, and I can't find the source, so I'm unclear what relationship Shandong Jiahong has or had to Leotek)
[1] 山東佳虹電子科技有限公司 is listed in this Lite-On document, as is Leotek: https://www.liteon.com/storage/document/document/file-ed4448... - or an English version of what I believe to be the same report, showing Leotek and Sandong Jiahong Electronics Technology Co., Ltd: https://www.liteon.com/storage/document/document/engfile-d21... Shandong/Sandong/山東 has gone as of the next financial report (2009): https://www.liteon.com/storage/document/document/engfile-d21...