US Streetlights Are Turning Purple
scientificamerican.com
scientificamerican.com
Acuity has acknowledged a phosphor defect in their lights and had launched a major warranty repair campaign, but I'm not sure how well that's gone given that new failures are still occurring. At least a year ago, they were struggling with the scale of the problem: it just takes a long time to schedule replacement of failed fixtures when there are so many of them.
If we were switching away from LED to sodium vapor lamps instead of the other way around, they would have written the exact same article but in reverse, still claiming the change makes us less safe.
According to letters to the editor in old newspapers, when major US cities switched TO sodium vapor lamps people complained incessantly.
Looking for references to "lights street yellow" in newspaper archives I also found:
* people complaining about the introduction of yellow traffic lights and how confusing and dangerous they were
* companies complaining about new FAA rules standardizing aircraft lighting
* people complaining about the introduction of street lights in general
People really only dislike two things: the way things are, and change.
When you're looking up in your rear view mirror, your view of the road ahead is from your peripheral vision. It most definitely does matter.
Instead of going bright white, high-color-rendition, wake-up-and-kiss-your-sleep-schedule-goodbye, which is what they went with instead.
When the LED streetlights were first installed, they were horrible, and we were promised they would be adjusted, but they never were sufficiently.
Maybe caring about the look of neighborhoods, and the sleep of people who walk or have unfortunate bedroom window positions, would've had a side effect of avoiding the problem in the article?
1800K white can render colors surprisingly well depending on the phosphor mix. I recently put one with claimed (and measured) CRI over 90 into a flashlight and was surprised to see that it actually can render blues reasonably well.
I'm inclined to think those are better choices for street lights than anything daylight-ish, but I also think we should use far fewer street lights. Their presence often reduces the contrast car headlights provide, making it harder to spot hazards while driving.
I’m not disagreeing with you, I agree that a high cri 1800k would be a nice night light. I just recently deep dived into this last week when my kitchen lights all died last week
We don't really need to see shorter wavelength colours like blues to avoid hazards. Night driving is not photography.
We could make a lot of improvements by rethinking our approach to street lights. When I've looked for studies on driving safety in the past, it did not appear that street lights improved safety in most places. Where they did help is crosswalks and low-visibility hazards.
At a controlled intersection, it would be easy to have lighting that's activated by a pedestrian pressing a button. Once drivers got used to that, the light being on would serve as a strong indication a pedestrian is definitely present at the intersection, even if they're not currently visible to the driver. For fixed hazards, small marker lights might be just as effective as overhead flood lights; passive reflectors might even have an equal effect, though I haven't read any studies on that.
This can easily be seen by looking at the shadows under an LED streetlamp, compared to a metal halide one. The LED shadows are very sharp.
> One thing that Bullough suggests pedestrians and drivers do to stay safe under purple streetlights—or any lights, for that matter—is to remove sunglasses and blue-light-filtering glasses when walking or driving at night.
https://www.irishtimes.com/ireland/dublin/2025/02/05/almost-...
I had someone complain to me about how it was a change made by the Green Party for bats: how ridiculous!
No, it’s just a faulty light…
I first experienced that in a bathroom in England. Door open, regular white light. Lock the door, blue light. I thought I broke something when that happened.
Though linking it to the lock isn't terribly useful, most heavy addicts won't care if anyone opens the door on them.
Makes you appreciate how tricky it is to balance cost, lifespan, and quality when you’re manufacturing millions of these for cities.
Pretty interesting.
Though that said, just go and replace them as they would have had to for sodium vapour lamps they had before? And this time replace them with something that runs them at a lower temperature, especially if the environment during a summer could be hot from external temperature.
There aren't any of them in my town, though I've seen one LED streetlight so far that went out. The main streets have lights with a slight yellow tint, and the side streets are neutral white, which I find much more pleasing than the purplish mercury arcs they replaced.
What an annoying/bad article. "Here are our guesses of this when we could have actually figured it out". It's not like these are in space and hard to get to, they are on the freaking street. Get a crew out there and figure it out.
Then they go on to do a _bunch_ of handwavy "science" about blue light while not really making any point (IMHO).
> One thing that Bullough suggests pedestrians and drivers do to stay safe under purple streetlights—or any lights, for that matter—is to remove sunglasses and blue-light-filtering glasses when walking or driving at night.
Ahh yes, for all the people that wear their sunglasses at night, I'll make sure to let Corey Hart know.
This just seems like an incredibly low-effort article with zero definite facts and enough hand waving to sprain your wrist.
Without them halogen headlights in my side mirrors give me a migraine after a while.
At first, I didn’t like it, but quickly grew to quite like it. (It’s pretty much only annoying for seeing the curb while parallel parking at night.)
Halogen headlights have been the most common (and relatively cheap!) kind on the road for many decades. They're just like the tungsten filament incandescent space heaters we used to use at home, except they burn at a somewhat higher color temperature. (And the way they work is really really neat, but I digress.)
Halogens, while higher-temperature than regular tungsten headlights (which haven't really been used at all in many decades) are generally "yellow"-ish compared to modern HID (aka "xenon" or "metal halide") or LED headlights.
Anyway, I'm also bothered by lights when driving at night, and I find the yellower corner of the spectrum to be the least-bothersome of common lighting colors. The glaring ~6000k white of modern HID and LED is much worse, for me, apparently because of the extra blue spectrum.
But if it really is yellow-ish halogen lights in your side mirrors that bother you most*, then I can't tell you that you're wrong.
And if that is then case, I can potentially offer some constructive advice: I've owned a couple of cars (specifically, a 4th gen Firebird and an E36 BMW -- both products of the middle 1990s) that came from the factory with side mirror glass that was tinted a pale blue.
The slightly-blue tint attenuated yellow halogen lights, by design. It's a clever bit of optical filtering.
During the day, in the sun where there's tons of light, they worked mostly-normally: Reflected images had little bit of blue tinge, but whatever.
And at night, the halogen headlights that were nearly-ubiquitous in the 1990s had their reflected intensity turned down automatically. Compared to cars with mirrors that used clear glass, I could still see the lights of the cars behind me just fine. They were simply less-blinding.
It worked great around the times those cars were produced.
If that's really the problem you're experiencing*, then a very pale blue window tint on the mirrors may be exactly what you want for that issue.
* (I asked for specificity because it's important. That Firebird and E36 both became increasingly-annoying for me to drive at night as HID and LED lights became increasingly-common, and towards the end of my time driving them I was seriously considering having clear glass mirrors cut and swapping over to that. Not because the blue tint wasn't effective at filtering out halogen light (it was great at that!), but because it seemed to magnify the problems I experience with HID and LED headlights. The world was changing, and the unchanged spectral filtering became a burden instead of a boon.)
The things that give me a migraine look almost purple, so I'm guessing it's HID / LED and not halogen.
I've found that yellow-tinted glasses help a bunch with those types, and I keep some in the car that I do use occasionally.
(There's a grossly misinterpreted study that people sometimes point at when saying that yellow glasses can't help and can only hurt, but the study itself doesn't support that conclusion.)
Deductive reasoning points to a failure of the phosphor layer in a specific type of led.
I mean, to be fair, the city maintenance crews don't generally run tear-down youtube channels with appropriate equipment to take a close look at a failing semiconductor-based light assembly.
But the general point is correct. What could be done is connect with some of those crews' offices, understand what warranties are provided with the municipal purchase of those bulbs, and how much comes out of the infrastructure budget to do an early replacement of these bulbs. Hopefully these bulbs weren't purchased from a RANDOMSYLLABLE amazon china dropshipper who disappeared after 3 months.
But yes, that requires talking to municipal governments, and there's not enough click revenue to support that level of journalism anymore.
Maybe a youtuber who runs a tear-down channel can do that.
This is how most LED bulbs in use today are purchased. In theory, LED is absolutely superior when engineered correctly, but it rarely is on a statistical basis when looking at the available products.
It's hard to make an incandescent bulb that is shitty, other than making it not last for a very long time. I'd rather a dead bulb than one that turns my street into a nightclub.
Since we had to switch to LEDs, I had to change lightbulbs waaay more often. I realize they consume less (even if here they would just help heat the house so the energy is not lost), but hell they are crap. They start blinking, their color is shifting, they die quickly, they are super expensive, they are bigger and many lamp I had couldn't be fitted with LEDs unless I found some smaller ones which are even more expensive.
This is my personal anecdotal experience, but I would be interested if any serious study had looked into this kind of issues.
But it's also important that leds receive the right amount of power. If they receive too much, they'll burn out. We've had this issue with a system where you clip little lights on a rail. It uses GU10 spots which exist in led, so thinking that was compatible, I used leds of course. But our spots kept burning out at an alarming rate. Turns out we need a different transformer for leds. I'm using halogen now, and it works fine.
I have switched all the lighting in 2 apartments to LED lamps, more than 10 years ago, and no lamp has gone defective until now.
However, they were Philips lamps and not one of their cheapest models, but some model with 1521 lumen @ 13 W and 4000 K color temperature, i.e. slightly more luminous than the classic 100 W incandescent lamps and with almost white light, only very slightly yellowish, very unlike the yellow lamps with a color temperature under 3000 K, and also very unlike the cheap bluish lamps with a 6500 K color temperature.
My LED lamps are screwed in traditional incandescent lamp fixtures, which hang from the ceiling, but unlike some bad lamp fixtures, mine have below them a diffusive screen, to avoid direct light, but they are completely open above, so they do not impede cooling.
I also suspect that many LEDs are designed for 220V, but here in switzerland it is 240V norm and in the middle of the day my power meter reads around 245V (I suspect it is because there is a lot of solar panels on the roofs around me), so it is not a negligible voltage difference if the electronic is not designed for it.
On the other hand my entire house is 100% LED from a prior owner gut and rebuild in 2015. With the exception of two LEDs the rest of them still work fine after over 10 years of use.
I've (not seriously) considered buying a pellet gun to shoot out the 4 massive neon purple lights at the entrance to my quaint 1970s era neighborhood. They didn't remove the old light poles after installing the new ones about 3 years ago, so it's double lit with sodium vapor and purple now.
Curious is there is a single person on the planet that prefers the white (er, purple) street lights?
But driver safety is probably way more important than my hobby
Yes
One at least, me
(I've never seen this failure mode, so I mean the LEDs)
I especially like the reduction in light pollution
I'd guess about 20% of the lights in and around Orlando are purple now. Maybe the heat and UV makes it worse here?
Also they are better suited to their job so need less light.
Sodium lights are nowhere near as flexible so are worse, mostly, on both counts
As far as directional lighting patterns go, that's a matter of the lamp design. Sodium vapor could be housed in a reflector that directs nearly all of it's light downwards.
I'm not against LED technology being used though, just the rampant failing LEDs and the obnoxious brightness & color they produce. Also not fond of the futuristic looking light pole design they seem to always use.
For astrophotography, sodium vapor's color cast is very easy to filter out, whereas the broad spectrum LEDs are not, especially when they don't all emit the same frequencies.
https://inside.lighting/news/24-05/heres-why-led-streetlight...
They don’t support this claim about led’s, and many groups are concerned about harsh cool light interrupting circadian rhythms. They’re also hideous.
Most LED streetlights replaced sodium vapor lights, though, which produce the sickliest, most horrible orange color known to humanity. Just about any LED is an improvement over those.
There were articles a few years back stating that the blue emissions from these LEDs were rather energetic and damaging to the retina. Conversely, some articles used to claim that red light actually improves the health of the retina. I don't know if those results were corroborated or debunked afterwards.
I know that personal beliefs and biases affect our perceptions. But such diametrically opposite experiences are surprising. I'm curious to know what everyone else experiences and any insights on this.
It's hard to think of a more normal sense of the word "orange" than "emitting and/or reflecting predominantly wavelengths between 590 and 620 nm." I guess you could argue that sodium is close enough to that lower edge to be yellow?
Low-pressure sodium lamps emit a single wavelength and they are the only kind of lamp that does not use LEDs, but which can match or exceed the energy efficiency of LED lamps. However, with low-pressure sodium lamps you cannot perceive any color.
There are also high-pressure sodium lamps. They emit a broad-spectrum light, even if with an excess of orange-yellow light. You can perceive the colors of things with such lamps, even if not very well. However the high-pressure sodium lamps have a much lower energy efficiency than LED lamps.
In Europe I have encountered mostly, or perhaps only, high-pressure LED lamps used for public lighting. I have used at home some low-pressure sodium lamps for certain purposes, but I am not sure if I have ever seen one like that used in a public space, here in Europe.
Low-pressure sodium lamps typically use transparent glass bulbs, like incandescent lamps or any other kind of low-pressure gas-discharge lamps, e.g. neon lamps.
High-pressure sodium lamps use special bulbs made of translucent alumina ceramic, because glass would not survive in those conditions.
A mixture of light with different wavelengths that is perceived as orange cannot be distinguished from a mixture of some pure orange with a certain amount of white light.
So any orange, of a flower or of anything else, has the hue of a single wavelength, but it may be more or less saturated, appearing like light with a single wavelength mixed with some white light.
Low-pressure sodium lamps emit a pure color that belongs to the yellow-orange range, so you could describe it as a yellowish orange.
High-pressure sodium lamps have a desaturated orange color, i.e. light that looks like a mixture of orange and white lights.
The orange of any kind of sodium lamp is much more yellowish than the reddish orange of neon lamps with cathodic light, like those used in neon indicators.
Most sodium lamps contain some neon for starting, so when they are switched on they may emit a reddish orange light for a short time, then change to a yellowish orange light, when the sodium vapor takes over from neon.
This is true but irrelevant when cost reduction is the motivating factor for switching to LED lighting, because that motivation will extend to the upfront purchase cost of the lamps and they will buy whatever is cheapest.
Unfortunately, most pole lights are 70CRI, too bright, and the light is too white (4000K+).
The flicker is pretty annoying because the transition is an abrupt on-off. Where I am the city had the bright idea to wrap LED ribbons around the poles in the downtown area to make it look more interesting. They connected the ribbons without diffusers directly to the power source of the streetlight. So what has happened now is that as you drive or walk you are looking directly at an unshielded flickering LED.
I still continue to install incandescent bulbs. They look better, and as I live in a cold-weather country the heat they generate is welcomed.
Flicker is also not a given, but a product of using cheap rather than good methods to power LEDs starting from AC mains power.
Most everyone has seen these now, in "Edison Bulbs" or elsewhere.
It would be nice to read some discussion of the core issue - the silicone - to me, it seems like a total failure of the silicone and not just an adhesion issue, as the photos seem to demonstrate small cracks leading to larger cracks and then total failure. Perhaps related to the addition of something new to the phosphor mix, which might change its properties under thermal cycling? Poorly mixed silicone? Unsteady temps as the silicone was applied?
The comments here!
Are we really such neo-phobe curmudgeons? Sodium hallide street lights preferred over LEDs?
Golly Hacker News, get a grip and join the modern world!
Also, I personally think spirited driving on a purple-lit nearly empty highway at night is a uniquely fun experience with an 80s-retro-futuristic vibe. The streetlights are merely accent lighting at that point; the highbeams are for actual visibility.