And then how do I know that they stick with the high quality approach? What happens when a brand decides to rest on the laurels of their brand name and start slipping in lower quality parts?
I follow his project a bit and it looks like consumers are really at loss. Generally there is no reliable way to choose a good led lamp without consulting such catalog. Lamps packaging often lies about actual specs, lamps with the same packaging but manufactured in different years might have different quality etc
A standard tag for Algolia/web search could work, if any spam comments with the same tag were flagged.
Instead of dumb capitalism, clickbait and silly content marketing driving human activity, to have an AI that saves us from wasting our time figuring out the answer to questions that have been answered long before. And then, instead, points us to those questions that have not yet been asked, much less answered. What experiments have not yet been done.
After all, no GPT-4, no matter how many billions of parameters, could tell you what the best battery is in the world if it wasn’t for that one human dude in his garage in Denmark, or Latvia, or wherever, who actually tested them all.
(random indie site: wet/dry shop vacuums, https://news.ycombinator.com/item?id=32411206)
The issue is that LED bulbs aren’t simple devices like incandescent bulbs. LED bulbs have an electronic power supply inside which drives the LEDs at a constant current.
Power supply design is a major subfield of electronics engineering and there are all kinds of tradeoffs you can make to optimize for different goals. Consumer electronics almost always optimizes for cost, to the detriment of all else.
It is possible (and not very difficult) to design LED bulbs that will practically outlive their owners [1]. The problem is that it requires putting more LEDs in the bulb and driving them at lower current. This makes the bulb cost more and the only benefit is longer life. For a manufacturer, there are nothing but downsides to this approach.
It is also possible (and not very difficult) to design incandescent bulbs that will outlive their owners. In fact, the first mass produced light bulbs generally lasted 2,500+ hours. In the 1920s, the major bulb manufacturers formed the 'Pheobus Cartel' in Geneva and secretly colluded to limit the lifespan of bulbs to 1,000 hours to boost sales [1]. Another example of planned obsolescence harming consumers and the environment.
[1]https://interestingengineering.com/science/everlasting-light...
Veritasium - This is why we can't have nice things - https://youtu.be/j5v8D-alAKE
The Great Lightbulb Conspiracy https://spectrum.ieee.org/the-great-lightbulb-conspiracy
> On 23 December 1924, a group of leading international businessmen gathered in Geneva for a meeting that would alter the world for decades to come. Present were top representatives from all the major lightbulb manufacturers, including Germany’s Osram, the Netherlands’ Philips, France’s Compagnie des Lampes, and the United States’ General Electric. As revelers hung Christmas lights elsewhere in the city, the group founded the Phoebus cartel, a supervisory body that would carve up the worldwide incandescent lightbulb market, with each national and regional zone assigned its own manufacturers and production quotas. It was the first cartel in history to enjoy a truly global reach.
You can dim them, and provide a slow start to prevent the inrush current (which is like 10times more than nominal with tungsten resistance increasing due so high 2500K temps).
[1] https://news.mit.edu/2016/nanophotonic-incandescent-light-bu...
That is a fantastic point. If they can't revolutionizing lighting in 7 years, it will never happen. Oh, btw, LED was invented in 1962, but it only took about half a century for the commercial viability of high powered LED lighting to appear and begin to take over the lighting market in 2011.
That said ...
... LEDs involve finding materials which exhibit specific quantum behaviours which correspond to human visual acuity.
Photonic bulbs involve the much simpler blackbody radiation concepts of not only incandescent light bulbs but hundreds to thousands of millennia of previous experience with combustion-based lighting ... and, yes, the added twist of finding viable IR-reflective / visual-spectrum emissive materials.
The second problem seems more reasonably simpler. One would hope that progress might be occuring at a more rapid rate.
(There's a similar argument I've used to contrast nuclear fission, which was commercially exploited within two decades of first demonstration, and nuclear fusion, which coming on a century from its theoretical understanding remains not even experimentally demonstrable on a continuous, energy-positive basis, let alone in commercial application. Some problems are just hard.
Yes, there are thresholds and breakthroughs, and they do occur. But given a few decades of lived experience matching advertisement to delivery, as well as a stronger awareness of historical examples and trends, patterns do become evident.
And that said: I will keep an eye on this. It does have the advantages of being simple, based on very well-proved technology, and a reasonable extension of same.
LED has severe problems, some solvable, some not. The article mentions flicker, but really should have specified. Nearly all LED drivers employ PWM. There are constant current LED drivers, and they are more efficient than PWM drivers, but PWM drivers are cheaper to design and manufacture. While many will claim PWM doesn't bother them because they can't detect it, they're not only exhibiting callousness for those that are bothered and harmed by PWM, they're falling into a fallacious trap, i.e. what they don't know and can't detect can't harm them, which is patently false, and one counter example is carbon monoxide. PWM LED drivers are now ubiquitous, and its effects range from annoying to painful, as anyone that has experienced migraine can attest.
Regarding the actual light LED produces, nearly all LED available are weighted towards the blue spectrum, and this light has been shown to massively mess with wildlife and shorten human lifespans by years by disrupting circadian rhythm, which strangely can lead to diabetes and heart disease.
LED proponents are obsessed with brightness, but this is also a trap, because brightness is not as important as what can be seen. Consensus among lighting and eye experts is that more can be seen with a dimmer light that reproduces color perfectly than with a much brighter light that does not.
As these issues with LED are mitigated, the drivers become more expensive and the LEDs become less efficient. And as the article mentions, the phosphors of better color-producing LEDs will fade rather quickly and over time no longer reproduce colors accurately. And let's realize that the best LED can ever achieve, what the technology has always been striving for, is to perfectly match what incan does with little to no development. Maybe someday LED light will perfectly match incan light, but it is not today and it isn't next year.
What it starts to look like, if the trend can be detected, is that as we fully mitigate the problems of LED, the more expensive it becomes until its cost and durability nearly reaches parity with what we already have with incandescent. It's starting to become a wash.
So unlike LED, if incan can be made more efficient, and there is a massive amount of room for improvement in efficiency there, then incandescent undoubtedly will return and dominate the lighting market. I liberally estimated 20 years, but it may take longer, but it really doesn't matter to my major point, which is that incandescent is coming back, and this is a very very good thing, because LED light, as efficient as it is, still absolutely sucks.
I'd pattern the inner surface of the glass envelope with a cube texture - think of taking a cube and pressing a corner normally into a clay surface, then removing the cube. This pattern is a so-called corner reflector, and returns incident light to its source. Figure the cube indentations at about 0.5mm deep, close packed. I'd deposit a dielectric film reflector stack tuned to reflect most infrared radiation onto this surface.
This combination would transmit visible light, but would reflect IR directly back to the filament, reducing the amount of electrical power needed to maintain filament temperature. Glass textural molding and dielectric film deposition are mature technologies. I think this could readily triple incandescent lamp power efficiency, maybe even better.
When we're talking market price, we have to acknowledge that it is a meeting of the price needed to bring a product to market and the price the consumer is willing to pay. We can't assume that the price of the longer lasting bulb would have been attractive to consumers, when compared to the price of the shorter-lived bulb, even if they had all the information available.
It's perfectly valid for a person to decide they'll spend more over the long run, rather than ponying up a larger sum now. And it's perfectly valid for producers to take the chance of deciding this for the consumer. As Henry Ford noted, "If I had asked people what they wanted, they would have said faster horses."
Was anybody stopping anyone from offering the consumer a higher-priced and longer-lasting bulb?
I retrofitted entire house - 200+ bulbs and fixture retrofits more that 5 years ago. I had one or failures since. I bought highest CRI bulbs, i.e. most expensive, and they work well. (Also, do not use bulbs for downlights - get entire "led can light fixture retrofit")
I used 1000bulbs.com because I can filter/read specs there, but you can get specs for any high-tier vendor and buy elsewhere.
How trustworthy is that state though? I have always assumed 0% trustworthy. Is that incorrect?
In general it's quite reasonable. Cheaper bulbs failed on me, more expensive ones work just fine for many years.
One thing is I wonder about if phosphorus (or whatever chemical they use) is burning out over the years. I.d. do I get worse light quality as these bulbs age?
Of course they won't be as efficient, and output very little light, but they will last forever. (also, no white, but hey, win some lose some!)
It's usually not that much, but I can see why they might eventually be upset about it.
Not all of them! I was very surprised to open up my generic outdoor patio LED bulbs and find two strips of LED filament wired directly to power.
AFAICT it’s just enough LEDs in serial for 120VAC at 60 Hz to be “good enough” that they survive for “long enough”.
To expand your metaphor: it's like judging a dish solely on the balance of flavors while not noticing that the restaurants have started changing the smell of the air, the firmness of the seats, the relative humidity and temperature individually.
At this point I believe companies are willfully refusing to inform their customers.
In other words, what they are selling isn't the same as the thing in the box nor the aggregate of all the components (since they interact with one another).
High end museums for example buy bulbs guaranteed to meet their datasheet specs, at a much higher cost of course.
I feel like I can't just have a casual fun hobby anymore. You have to have all of the knowledge about the entire space just to be able to decide if something may or may not be garage.
This has always been the case. The difference now is that with the internet it's within reach.
You don't have to dig through your social network to find someone working for the lighting division of GE. You don't have to visit your local library to check out books on how lightbulbs work in order to figure out which makes one better than another. You just need to hop on Google or ask New Bing.
--
Think all incandescent bulbs were the same? Think again. Manufacturing conditions and filament thickness are two of the several factors involved in how long that lightbulb will last and how bright it will get. Cheap, shitty lightbulbs from discount stores were a thing.
Oh, and one more thing! You're pretty much stuck with one color temperature.
--
There are plenty of examples throughout the 20th century of poorly-made, barely-working tech being sold as acceptable. The plethora of non-electric "vacuum cleaners" sold around the turn of the century are one notable early example. The lightbulbs which came after the agreements made by the Phoebus Cartel are another.
1978! Home video! Do you go with VHS from JVC, Betamax from Sony, SelectaVision from RCA, or DiscoVision from MCA?
For an entertaining diversion, imagine you're living in 1973 and it's time to purchase a new car. Is that Plymouth really going to hold up against your new concerns about gas mileage? How do you know? Do you have any mechanic friends? Do you know anything about how cars work? Does the local library have any books to help?
Random final tidbit: The "older"=="better" myth is the result of the fact that we're not exposed to the junk of yesteryear; only the good stuff. The junk was thrown away years and years ago.
Awful junk existed before the internet. It’s just that people didn’t have much of a way to know any better, nor did they have many options to choose from. People relied on word of mouth, marketing material, or the shop keeper’s advice to decide what to buy… that is, if the store even had multiple options.
There’s not more crap today, there’s more perspective.
- Efficiency >= 45 lumens per watt
- CRI >= 90
- R9 Color Rendering value >= 50
- Rated life >= 15000 hours
- Minimum dimming level <= 10%
- Flicker <= 30%
Theoretically if a bulb is listed as JA8 compliant (and the certification isn't fake) you know it at least meets these thresholds.
And what do you do when the lightbulb burns out after only 3 years? The product has long since changed SKU, the manufacturer gets to claim they fixed any deficiencies (and it'll take years to find out if they are telling the truth), and you long since lost any proof of purchase.
https://energycodeace.com/site/custom/public/reference-ace-2...
This one was a little tricky for me when I was buying bulbs last year. I prefer warm-colored bulbs, and I was kind of confused why Amazon kept on saying it was refusing to ship bulbs to me. It took me a while before I realized it was because I'm in CA and the CRI was too low, and Amazon didn't have a way to just filter by CRI. Eventually my wife just ended up finding some warm-ish LED bulbs at a local store.
Like any informed consumer you must read every retail-based HN comment thread ;-]
6 of them have failed in less than 9 months, either flickering so badly it could cause an epileptic seizure or just straight up dying on me.
It's maddening.
It's possible the quality has changed, but i'm also wondering whether the mains voltage might be a factor - there is quite a wide range of possible voltages allowed whilst still being in-spec, so maybe i'm lucky at my properties and mains voltage is on the low end of the standard and maybe you're running hot. It's all most frustrating!
BTW, I went with Philips on the basis that there was a good chance that if I did need to replace a few after a year or two due to failures i'd be likely to be able to source the same bulb, as it's really annoying if you find one bulb a different colour than the others...
With Philips spinning off the lights division, I don't know the current quality.
On the other hand, I've had others (non smart ones) starting to flicker or otherwise just dying after as little as one year.
PITA I know.
Consumer Reports tests a lot of consumer goods and used to be my go to for testing. They don’t take ad revenue so that helps. Though you have to be a member to see their reviews.
https://www.consumerreports.org/cro/about-us/what-we-do/inde...
They used to be the go to for car reliability info (when your car was rated high you sold more, according to my dad who was in the car business)
I can't edit my post. Trusted reviews are hard to find, I would still trust Consumer Reports over random Amazon reviews.
There is simple heuristic - all LED bulbs are bad. They generally have insufficient heatsink and unreplaceable PSU immediately next to LEDs.
Better are LED tubes (with T12 interface, as a replacement for fluorescent tubes), they have much more area for cooling and for PSU, and sometimes have replaceable PSU. Similarly lighting units with integrated LEDs.
At the end of the day, the author recommends giving filament-style LED bulbs a try, where the emitter is built onto a thin filament away from the circuit housing so it's far away from the heat.
A scale that can measure grams is like $10.
That's the million-dollar question of the '10s and '20s (and likely beyond), and is far bigger than just bulbs.
The only solution that I can think of that could work is a distributed rating system with a built-in web of trust. It theoretically shouldn't be that difficult for people to adopt, if only they got collectively fed up with the universe of crap we have now, and someone provided a nice app and protocol to federate information with.
(direct regulation of quality, vendor-controlled ratings, and browsing Reddit/HN comment threads are all fatally flawed non-solutions)
On the other hand, Daniel Kahneman was awarded a Nobel prize in 2002 for researching with Amos Tversky on how we make decisions, and how having more options makes our eventual decision less fulfilling as we suspect that we probably did not make the optimal decision. However, done is better than perfect. At least that's what some people say.
Given that we have multiple technology purchases to make, all of which will involve "research" and making decisions it is very frustrating, to me, that we do not have more reliable trustworthy guidance. There are competent review organizations and websites but they more frequently tend to be owned by product manufacturers and funded by advertisers. We know that marketing tries to create desire in our primitive consumer brains.
And as individuals with deep and long experience in at least one or more areas, we have our own biases that help us make decisions. And if we think carefully about how we gained this expertise, we should conclude that a lot of wasted time and mistakes were involved.
And we know that becoming an expert in lighting, spectral, power consumption, lifetime, CRI, etc could take a long time and there would be more to learn as the engineers create new solutions (blue LED plus yellow phosphor, or RGB LEDs, COB or something else...).
So to answer your question. You won't know that brand X produces quality bulbs that last a long time until you purchase and test them. Assume that they won't last a long time, don't buy the most expensive option, there will be improvements in LEDs and bulbs that will make your next purchase even better.
To sift past the marketing to get actual quality, well we do have some well known brands that distribute through well known stores. Buying from Alibaba or the dollar store is not going to result in the best outcome, but it might. Put those options aside as an experiment rather a "must make me happy now" experience.
How do you know that a product won't slip in quality over time? Well you won't know until you make that purchase. This happens all of the time with everything from salt ("Himalayan" salt with rocks, sea salt with microplastics, honey adulterated with sugar, olive oil with other oils)...
We live in a very interesting time, we no longer have to "follow the herd" or "hunt for the roots" in new locations. We're mostly protected from weather, earthquakes, famine, etc (exception occur). Our health generally good. There is however dog poop on the sidewalk and pot holes in the roads.
So when the grocery store moves your familiar product to another aisle, or changes their product line up, or increases the price, these are all opportunities to step out of "cruise control" and experience the uncertainty that comes with a constantly changing world.
"Marriage has no guarantees. If that's what you're looking for, go live with a car battery." - Erma Bombeck
Up to 70 years ago, item X had an extremely low complexity, and up to some 30 years ago, supplier Y had reliably constant quality between its products.
No previous generation had to deal with the problem we currently have.