What Happened to the 100000 Hour LED Bulbs?
hackaday.com
hackaday.com
I film all the LED bulbs I buy in slow motion at 240 fps with an Android phone, and play the movie back on a PC with mplayer using the dot(".") key to move frame by frame. Here is a shot I made comparing 2 brands of LED bulbs: https://youtu.be/QbenId_F2RQ (Edit: yeah you don't have to transfer to a PC, just playing back in slow motion on the phone will still show the effect quite well.)
It's amazing how I find this way that most (but not all!) of LED bulbs flicker with a strobe effect at 120 Hz (frames alternate between bright and dim) because they have crappy power supply designs that fail to smooth the A/C voltage. As a result they flash one time during the positive phase and one time during the negative phase of the 60 Hz A/C mains frequency.
I find this unacceptable. Although not too noticeable in normal conditions, a 120 Hz strobing light is definitely noticeable when your eyes move or track an object illuminated by the bulb.
In my experience, Philips lightbulbs are one of the few brands that don't have this flaw because they take care of converting AC to a stable DC voltage internally. In fact they are advertised as such: https://www.amazon.com/dp/B07CFRCGKC "COMFORTABLE LIGHT: Our products meet strict test criteria including flicker, strobe, glare and color rendition to ensure they meet EyeComfort requirements"
Of course it's still hit or miss, compatibility between bulbs and dimmers seems to be very hard to judge without testing these days.
Most dimmers work via pulse width modulation [1], i.e. the power supply is switched on/off in a different interval (more off time means lower brightness). A true flicker-free dimmed LED has to be dimmed in an analog way (probably only feasible for "dumb" light bulbs). PWM of 20kHz or something might also work, but who knows what that does to brain waves.
1: https://www.waveformlighting.com/film-photography/an-introdu...
Visible light impulses influence brain wave patterns (e.g. [1]), and it might not be beneficial to look at pulsed light with a frequency that is not well researched.
Same thing with screens, most LEDs are dimmed with pulse width modulation [2], so they flicker at most brightness levels except the very brightest one. Some iPhone X users claim they get a headache when looking at the phone at the lowest brightness setting (OLED is even worse in that regard than LED, IIRC).
1: https://clinicaltrials.gov/ct2/show/NCT03657745 2: https://www.eetimes.com/document.asp?doc_id=1281013
On iOS there’s a setting buried deeply in accessibility settings, I think it’s called white point correction.
And on iPhone XS this is even worst. I wonder if there are any solution to PWM problem.
Especially when my dog bounces around it looks like she is in a strobe light. I did a couple of videos of it just to "prove" it in some fashion.
――――――
The worst bulbs I've personally seen for flicker were dimmable ones.
The non-dimmable variant from the same range had no noticable flicker.
Price paid also seems to be no indicator of quality – I've had premium LED bulbs from major makes, bought from a reputable wholesaler, that have died after 6 months within days of each other.
I've had cheap no-name imports that have lasted years and had no flicker.
However, there are also LEDs that support legacy dimming, but again translate less current into different flicker patterns...
In other places I don't have dimmers but have low-voltage (12V) cans. Lots of flicker issues there too, presumably due to the 120->12V transformer at the can.
And then I have a light fixture that's both dimmed and a custom 12V setup. I've given up on finding a working dimmer/LED bulb (MR16 GU5.3) there. Probably need new 12V transformers?
Here are two good sites that posted graphs of light intensity over time for a bunch of different bulbs: http://www.ledbenchmark.com/ http://sle.se/michael/led/ (Swedish)
Unfortunately, they both stopped updating around 2015. Anyone know of similar sources for more recent bulbs?
> “I am on a mission to make lighting healthy. That's why I made Bedtime Bulb: https://bedtimebulb.com/ It has the lowest flicker I've ever seen in a bulb form factor—even less than those claiming to be "flicker-free."”
In addition to that, a few other database sites that test for flicker were mentioned here³.
――――――
¹ — https://news.ycombinator.com/item?id=18644590
I don't have a database, but I can say that most (but not all) filament designs are terrible for flicker--ours is an exception. Philips products seem to have gotten worse in quality over time, probably to remain competitive, but they are better than the worst. Ikea's stuff is pretty good overall. I don't have a consistent recommendation for low-flicker general lighting right now, as every brand has good and bad products.
And, yeah, I tried that Soraa healthy bulb, and it was shockingly bad, given how it was advertised. Even the regular Soraa bulbs were disappointingly flickery, though less so. Also, just looking at the pattern it makes on my phone's rolling shutter, the wave is quite saw-toothed, and I wonder if that makes it worse? I have some Ikea candelabra bulbs which were always pleasant, and I was surprised to see they also left quite a visible flicker pattern on my phone, though much more sinusoidal/smoother.
It's a shame about the flicker in the Soraa bulbs, since the colors really did seem nice.
What's interesting is that the bulb seems to have a nice build quality (from the outside, I didn't open it) and is quite heavy. I suspect they are putting a BIG heatsink in there, as violet pump phosphors tend to degrade quickly.
The other thing that irks me about this product is the "zero blue" messaging. Yes, I can confirm the spectrum looks like what they are advertising, with a peak around 415nm, almost nothing between 440-490nm, and then all the rest of the colors from green to red.
But, green light is just as bad as blue for sleep! The ipRGCs are sensitive to both blue and green (up to around 600nm) This bulb has a lot of green, more than a lot of other light sources. It also has a very unpleasant greenish color (IMO) and low-ish CRI, around 78 on the bulb I tested.
Is it better than a normal 600lm bulb for sleep? Yes. Is it novel? Yes. Is it good in practice? I don't think so.
what does that mean? Duty cycle is 50%, frequency at something Hz? Not being snarky - curious whether there is some kind of standard that your tool compares against.
A duty cycle 50% (or any percent really) isn't bad in itself, it's the frequency that is the culprit. Too low PWM freq and you'll notice it.
Two factors are often used to describe flicker for lighting:
- Modulation %, a.k.a. % Flicker [The height or modulation of the waveform, formula is 100% * (A - B) / (A + B)]
- Frequency [self-explanatory]
Flicker Index is also used (refer to [1] page 7 for more info) but rarely.
Duty cycle is not considered relevant, as you mentioned.
Some tools designed to measure lighting, such as my UPRtek CV600 [2], spit out % flicker, frequency, and flicker index, in addition to a bunch of spectrophotometry metrics. It's also pretty easy to calculate the modulation % with an oscilloscope and photodiode, it's just Vp-p / Vmax.
IEEE 1789 [0][1] roughly says that the higher the flicker frequency, the more acceptable it is to have a high modulation %. If you look at the graphic on [1] page 18, you can see this relationship. The white area is considered unsafe, the yellow area is "low-risk", and the green area is "safe." I don't 100% agree with this personally, as most incandescent lighting, with around 6-11% flicker at 100-120 Hz, would fall into the unsafe or low-risk category. But the Soraa product is definitely in the unsafe category.
For reference, I plotted my product (Bedtime Bulb) against an incandescent A19, Lighting Science's Goodnight A19, and Soraa's Healthy A19 on an IEEE 1789 graphic [3] with a Python tool I'm developing, Beautiful Flicker [4]. You can see that this incandescent, measured around 10%, is on the border between unsafe and low-risk by this standard.
Hope this helps!
[0] IEEE 1789: A new standard for evaluating flickering LEDs?: https://www.dial.de/en/blog/article/ieee-1789-a-new-standard...
[1] Flicker: Understanding the New Recommended Practice (PDF) https://www.energy.gov/sites/prod/files/2015/05/f22/miller%2...
[2] UPRtek CV600 https://www.uprtek.com/en/product/SPECTRAL-COLOR-METER/CV600...
[3] IEEE 1789 Graphic Comparing 4 Light Sources https://github.com/yeutterg/beautiful-flicker/blob/master/ou...
[4] Beautiful Flicker https://github.com/yeutterg/beautiful-flicker
This is slightly worrying.
The stroboscopic effect is really fun right until you're in a place with lots of moving parts at a multiple of the strobe frequency.
OTOH it's also used to fine-tune engine timing and such: set a stroboscope at a fraction of the frequency you want, the engine is properly tuned when everything looks completely stopped. Makes it very easy to notice mistiming.
I do some occasional wood-turning, but usually augmented with ambient daylight - along with the dubious safety advantage of having a very noisy electric motor.
Plus I'm old enough to have used a timing light hanging off a distributor on my first car. Try explaining that to the kids of today. ; )
The flicker is a thing from the past (or should be).
* at least in the UK
The CRT was a 17” running at 60hz. The rubber band sailed past in the span of about 2 screen refreshes, so it traveled that distance in 2/60 of a second. Distance was easy to calculate by the Pythagorean theorem or by holding a ruler to the screen, I forget which.
I saw a test in a Swedish newspaper lately that showed IKEA's cheapest lightbulbs as one of the less flickery ones in the test. IKEA art.nr 303.887.64. So much for the price to guide you.
LED lights are alright. Lots of people think that everything used to be better but my memory is too good to fall for that trap.
The rotating / reciprocating mass may appear stationary at some RPM.
Just a reminder to only ever use incandescent or halogen lights around machinery you can touch while in operation: drills, lathes, mills, slotting machines, etc.
Not saying it can't happen but it does seem stupidly hard to achieve.
Even after 20 years in the metal fabrication industry that scenario still spooks me.
Edited to add: Have you ever heard of a timing light for an engine? This works exactly the way you are describing.
Never underestimate the ability of machinery to rapidly render you dead.
Never underestimate the ability of otherwise intelligent people to intentionally disable safeguards, or guninely make mistakes.
So far, the discussion sounds like it's a theoretical possibility. If it's a real possibility, then it's something that should have happened a number of times.
The timing light is then pointed at the flywheel on an engine, which has numbers or marks stamped into it. Each time the spark plug fires the timing light (which is hooked into that same current via induction) lights up for a brief amount of time to show at what timing offset the engine is currently at. (this all happens at hundreds of rpms a minute).
so I'm not sure it's the same
If the light is flickering it's absolutely guaranteed that it's derived from this.
But it's hard to imagine a lightbulb that emits a short enough pulse of light to make something look stopped. Even an absolute garbage one-way rectifier will be emitting light more than a quarter of the time. That can make a tool look odd, but it won't make it look still.
Not necessarily. If it's a 50/60 or 100/120 Hz flicker then yes, but LED lights with a cheap switching power supply might still flicker at the switching frequency, which could be say 400Hz or something.
Many AC motors are synchronized to the frequency of the power line, by the same principal that turbines in the power grid are. So 60hz AC often means 30hz = 1800 rpm motor, and 59.8hz AC means 29.9hz motor. So if you have LEDs with a half bridge rectifier and your machinery is 180 degree rotationally symmetric, it will appear to not be moving, even if the line frequency fluctuates.
Fluorescent tubes are most noticeable, LEDs a little less. Tungsten or Halogen are the only sensible option for workpiece illumination.
For my hobby stuff, I just bought a couple of spare halogens that should see me out.
Most will flicker at a multiply of the power line frequency (60 or 50 Hz depending where you live).
It's so common, that most digital cameras have a setting to set the power line frequency, so they can reduce the flickering in the footage.
Incidentally - many alternating current engines are also working with multiples of the powerline frequency, because it's easier that way.
The rotation frequency would have to be close to a relatively simple multiple (or fraction) of AC frequency but the way electric motors are contstructed if the piece is directly connected to a motor it is very likely it is true.
Before you start a tool you should act out your planned motion. Then start the tool and only perform this motion. If something changes and you cannot complete the motion, stop the tool immediately. Following this and a few other safety tips I've operated power tools for years without incident and I'm legally blind.
Okay, but, say you've configured the tool wrong and now it's destroying your work. You stop the tool. How soon is it safe to grab your work away from the work surface—i.e. when will it go from "it's too dangerous, let it keep destroying the work" to "it's now slow enough to pull the work off of the tool while it's still—slowly—running, to save the work from further damage"? That's a thing you have to figure out with your eyes (and ears).
So, evidently, I may have over stated my initial claim about only using incandescent lights.
Still, it’s something to me conscious of.
――――――
Of course it's not the LEDs that do this but the cheap AC-DC conversion bridge rectifier.
Through trial and error I found that 72hz was acceptable/ok and 90hz and above was great. Other folks generally didn't notice, but sometimes could if I showed them to look sideways etc. This must explain why I've never knowingly been bothered by 120hz LED flicker, it is below my ability to consciously notice.
Curious, I took a look at several bulbs, with the mobile at 240fps trick described elsewhere. A new Feit dimmable, new Ecosmart non, a several year old Ikea, and a no-name CFL. All showed flicker except the CFL. All the LEDs flickered, but the amplitude was larger on the ecosmart. Feit and Ikea had a small amplitude.
Perhaps they are putting half the LEDs out of phase to get the smaller amplitude, or they should?
I mean, it looks absolutely fine as long as the light source and viewer are both static - just a shame vehicles tend to be moving :(
The switch-mode power supplies in proper lamps get the stuff done way better and there will be no noticeable flicker. Maybe in the order of several kHz in ripple which you cannot see.
Also; cheap off brands (looking at you china) sometimes have caps rated at 250V which is far from the peak mains rate of about 320V (in 220/230V countries).
[1] https://www.powerelectronictips.com/whats-stuff-capacitive-p...
Lights is something that I used to cheap out on but no more. Not buying Philips bulbs is definitely 'penny wise, pound foolish'.
Also, Philips Hue!
Do I need to buy a specific model of Philips bulbs? Or is my Ikea lamp the culprit now? I paid a ton of money for a certified flicker free monitor so I'm very keen on making sure the lamps are flicker free as well.
Source: I just had a problem with a malfunctioning light switch (not even a dimming switch, a plain two-state switch) wherein sometimes it would poorly complete the circuit, and one bulb out of the four on the circuit would start oscillating on and off. (If you removed the one bulb, another would start. If you plugged the first bulb back in, in the original slot or another, it would happen to the first bulb only.)
Alas this isn't a problem. If a lightbulb goes, Amazon send me a new one out now free of charge and tell me to throw the old one in the trash.
This is one manufacturer, prevalent on Amazon, the Long Life Lamp Company. Long Life my ass.
They have been replaced by Philips and Ikea LED bulbs now which the oldest are 5 years old now and still going strong.
Edit: one thing to note is the really cheap ones run pretty hot. They have 105 oC rated capacitors in them. If you look at the derating curves at the running temperature they are clearly designed to last just past a year.
Worse than that, though, it was unusable with a dimmer switch; the whole thing visibly flickered every few seconds. I felt bad returning it because I really wanted it to work out, but the quality just wasn't there.
My other LED fixtures were ordered through aliexpress. The fixtures were fine, but the transformers that came with the fixtures all failed and had to be replaced (overheated, melted the plastic casing). This turned out to be tricky, because the LED drivers available on ebay and aliexpress are never the same, there are constantly new shoddy ones from noname brands, and the ones I had bought before tend to be nowhere to be found when I need a new one.
[1] Except after a power outage - then they eerily flash on and off like something out of Stranger Things.
The sad thing is that the latter is still cheaper than buying from a 'local' shop.
I also own some 5W/400lm ikea bulbs for general room lighting and they do run warm, maybe something like 40°C ± 5°C. The fixture/thingy around the bulb is roughly the size of a fist.
The bulbs are pretty new so I can't say anything about the lifespan of the bulbs, but the price is reasonable, they are not the fancy smart bulbs.
One thing to pay close attention to on the package is whether the bulb can be run in an enclosed fixture. The package should always specify. Following this guidance and picking the right bulb for the fixture will help ensure you get the service life you expect. I don't think I've found any 100W equivalents that are OK'd for fully enclosed application, and 60W equivalent is like 50/50.
Edit: looking at the price of fixtures, the economic argument is poor as well.
With a retrofit lamp, you can use a high quality trim baffle (these things cost about $12, are quite well designed, and are probably already in your house and put a civilized PAR30 or PAR38 LED in, and you get a much better light distribution.
Back when lamps were all standardized incandescents, fixtures could focus on quality. Now, with LEDs, I suspect that manufacturers focus on how cheaply they can build a driver and assembly that meets some minimum performance criteria. So, with retrofit lamps, at least the fixture design is separate.
(There are exceptions, of course. Very high quality dedicated fixtures exist, but they’re quite expensive.)
Kind of like how replacing a gas dryer "requires" a licensed plumber to make the gas connection or you're "required" to drive no faster than the speed limit at all times.
You're getting the runaround. If it was in a vacant apartment that they were showing to people it would already be fixed. If they cared about getting it done ASAP they'd either shell out the big bucks or ignore the law.
Edit: I think a lot of people on HN don't realize how simple swapping failed parts is. Swapping a failed water heater or electrical fixture (especially light fixtures) is a waste of the skilled tradesman's time and the customer's money. Even an incompetent landlord or maintenance service should be able to get it done.
IIRC, Australia requires an electrician for even this sort of minor work.
example: https://www.progressivecommercial.com/business-insurance/han...
That likely depends on local codes; not sure about light fixtures specifically, but I definitely have encountered variations in different cities in the same county when it comes to ceiling fans.
Those laws (especially when combined with licensing requirements that exclude all but those who have a full time career in that field) are a massive hindrance to individuals and many businesses. They usually amount to little more than a potential transfer of liability. If your house burns down maybe you can try to point the finger at the licensed professional (good luck at that). The only thing it accomplishes is making it (slightly) harder for people who are habitually grossly negligent to keep working in that field. This is a non-benefit IMO because so much business is based on reputation anyway. The downside is that by artificially raising the minimum price access to access skilled labor you cause all sorts of work to be forgone because people can't afford it.
Plenty of house fires have been started by people who couldn't afford to pay an electrician but could afford an extension cord.
Edit: Since I'm apparently so wrong does anyone want to explain why. A down-vote without a reply is roughly synonymous with "your opinion is inconvenient to me but I cannot refute it"
* Home rennovation (requiring city signoff)
* Home repairs
* Rental repairs
* Court over repairs or alterations to property
...is using the downvote correctly. Downvoting is for comments that are not productive contributions to the kind of discussion desired on HN, responding to any post that merits a downvote is increasing the noise-to-signal ratio.
Edit: I've seen this at least in NY, Florida, California
I've read that you can't move chairs, etc, either. I understand it's part of the contracts between venues and unions.
It's rarely criminal for homeowners to DIY within building code regulations, permits be damned. When the bank owns part, appraisal or home sales happen, insurance policies are involved non disclosure can be fraud.
My 2¢ on DIY from general contracting experience,
Water: Undetected water leaks can be fucking expensive, consider location and worst case scenario when making decisions. Don't use shit like AS SEEN ON TV toilet tank gaskets or tool free water valves in a production environment.
Gas: Thread seal tape is a lubricant that helps seat solid metal (usually) connections to create a seal, not a sealant. (can be welding agent for polymers)
Water, dish soap, spray bottle helps test connections.
Gases have different smells or lack there of, different densities (sink, rise, displace oxygen), can combust in different concentration ranges, etc so understand this.
Electric: Changing the load or adding potential for increased load on a circuit (new fixtures, receptacles, breaker, fuse, conductor, etc) could cause arcing and lead to fire in unexpected ways. Sections of copper wire get replaced with smaller gauge at some point over a properties history and then get sealed back up in a wall, among all kinds of other wacky shit. Understand the physics and dangers.
Connections are important, and common points of failure. They should be appropriately terminated and enclosed. Use the proper size wire nuts.
All: Valves, switches, and people are flaky so plan accordingly. Test shut-off or failsafe mechanisms and have a plan for the worst case scenario.
In other countries (both in Sweden where I lived before and here in Japan where I live now) there are special ceiling power outlets, so replacing a light fixture is just a question of unplugging and plugging in again. Japan: https://farm5.staticflickr.com/4279/35698610491_8787f45eae.j... https://farm5.staticflickr.com/4230/35790814576_382943d360.j... Sweden: https://www.eldirekt.se/pub_images/medium/1890776__13222.jpg
There's a wide range of quality ceiling light fixtures that you just clip into place, and the LED chips are on a huge metal heatsink so they run very cool https://panasonic.jp/light/led/products.html
They've run cheap and cold for years now, without a single failure.
I suspect that there's some planned obsolescence in newer products.
0: https://www.amazon.ca/gp/aw/d/B01EN9L8YU?psc=1&ref=ppx_pop_m...
I think it gets worse as time goes on, because manufacturers gain experience with the lifespan of their components and learn exactly where they can cut. If you’re not quite sure how long something will last, you’ll probably err on the side of caution and engineer it to last longer than needed in case you got it wrong.
However, as the article notes, people may not be interested in paying a premium for a product that lasts 10y instead of 5y or 20y instead of 10y, since the tech can change or maybe it’s susceptical to other kinds of degradation.
These were fairly expensive, about $11 each (the going rate at the time), and were TCP brand, but it was unpleasant to read that newer bulbs might be far worse when it comes time to replace them, even if they are cheaper. I love the fact that they die so infrequently that I don't have to keep any spares around.
EDIT: This was not entirely correct as the article says:
> The US Department of Energy (DoE)’s solid-state lighting program supports research and development of LED technologies, and their website contains volumes of data on LED lighting systems. Their Lifetime and Reliability Fact Sheet contains data on the failure rate of 5,400 outdoor lamps over 34 million hours of operation. Interestingly, the LEDs themselves account for only 10% of the failures; driver circuitry, on the other hand, was responsible almost 60% of the time. The remainder of failures were due to housing problems, which may not be as applicable for bulbs in indoor use.
Some don't even have a bridge rectifier - just a couple of capacitors and that's it.
The Walmart bulbs without any special pricing were way cheaper than the Cree bulbs from Home Depot, but even better, there was some sort of automatic rebate or something like that in cooperation with the local electric company that made the Walmart bulbs $0.17. By "automatic rebate", I mean that the bulbs rang up on checkout at $0.17. No rebate forms to send in or anything like that.
If I cared more I might try to take the bulbs apart in a way that I could get them back together and just bypass the failed LED with a resister to bring the bulb back to life (at 15% reduced capacity or so).
I have been very disappointed in how hot the bulbs run as well. It feels like they must be wasting a lot of energy to run that hot.
Could you tell us that email address ? I have a box of 14 original Cree bulbs that I would like to get replaced...
Were those the original Cree bulbs with the heavy, finned metal heatsink around the base ?
I currently have a bag of about 14 of those that I bought (with great enthusiasm) and that burned out (or turned weird purple colors) within 2-3 years.
New style cree bulbs appear to have these issues solved. In fact, I continue to buy them as they perform better than other (satco, fein) bulbs that I have. In most cases I get satco/fein as long as they have the color temperature and output I want, but if I am having issues with a dimmer, etc., I get a cree bulb.
It is a problem in the sense that it creates waste and pollution. There is a cost to throwing things out, and for shipping things across the country without limit.
I do agree with you fundamentally however.
the first being, areas with vibration meaning near doors and garage door openers. great locations to eat LED bulbs. this might be related to your heat sink adhesive issue.
the second is the upside down tulip light fixtures common in bathrooms, apparently they cook quite well unless ventilated at the top.
My flat has some sort of bad-quality switches and when switching on/off the lights I often used to have some failures when using classic light bulbs (I have no clue about electricity - in any case in some way when I press the on/off-button the transition seems to be "dirty").
4 months ago I bought a new lamp and decided to give OSRAM-bulbs a go (it's a well known brand in Europe but up until then I always used Philips) => 2 LED-bulbs failed within 2 weeks (I put in the 1st bulb => failed after 2 weeks => replaced it with a 2nd bulb => failed again after 2 weeks).
I then went back to using Philips-bulbs and the one I used to replace the last one which failed is still working now.
Philips seems to be, at least in my case, definitely better or at least more "lenient/tolerant" at least concerning my the switches installed in my flat.
Anybody having negative experiences with Philips or positive ones with OSRAM or the opposite?
It never answers the question. It has an anecdote about the author's own lamps that ran an estimated 15-20k hours with an advertised lifespan of 30k and have degraded to the point where they will replace them. That's n=1 and nowhere near 100k hours, or even 50k as the author mentions early-adopters will remember seeing advertised.
The article contains a lot of info, but does not actually answer the original questions it poses in the title and the second paragraph.
>It’s possible that the reduced lifetime ratings we see on current bulbs simply reflect better knowledge about actual performance of existing LED technology over time.
Supposedly that's better for night vision, light pollution and various animals. But everywhere I go I see yellowish sodium vapor lights replaced with white LEDs.
The lumens out of a legacy high pressure sodium lamp are very high. It is still an engineering challenge to get comparable lumens out of LEDs at price point the market will bear.
http://adsbit.harvard.edu/cgi-bin/nph-iarticle_query?2001IAU...
Source: https://www.amazon.com/LED-Lighting-Primer-Future/dp/1449334...
https://www.ies.org/product/projecting-long-term-lumen-maint...
But, in practice, other components generally fail before the LEDs have a (binary) failure.
- longevity - when used with a dimmer - when a large load is applied to the same circuit
Most other bulbs fail one of those, but I haven't had an ikea branded bulb do poor in any condition.
1) Better cooling: non-tooth paste thermal compound, more space for heatsinks
2) Not overdriven LEDs
3) Introduce DC power lines (within the house/building) - galvanic corrosion still might be an issue
4) In lack of point 3 - more efficient AC -> DC converters, preferably with lower temp threshold,
having 150C default for the internal mosfet on 8pin LED driver is really high.
4a) AC-DCs require quality passive components: coil and caps.Our eyes are most sensitive in the green, but green LEDs are not necessarily the most sensitive
A high efficiency LED assembly is still producing on the order of 50% light and 50% heat.
I know for a fact that some of the testing houses have done extended on/off cycle testing for some of the big box retailers looking to keep the lamp vendors honest on real lifetime performance of A19 class bulbs. I've been in such a test chamber.
I've tried through 5 different MR type LED bulbs and every one of them had an extreme flickering issue. All of them were labeled as dimmable.
I've had better luck with the standard A type bulbs. But only because there were more thorough reviews available.
If dimming is important to you, there are a number of architectural grade bulbs that offer deep dimming to 5% or 1%.
Personally, for this application, I've had good luck with Cree bulbs of this type at home (on Leviton dimmers). And in the house of worship environment with Cree bulbs on Doug Fleenor dimmers (http://www.dfd.com/dmx8dim.html)
I'd also like to add that looking for lamps with the CEC Title 20 & Title 24 certifications should be helpful in finding bulbs that dim well, as California has much stricter regulations on dimming flicker and buzz than the rest of the country.
The cheaper bulb in the original article has linear current control. Inside the ASIC, this could be as simple as a BJT or FET in saturation with a (somewhat) controlled base/gate voltage, maybe with some temperature compensation.
The linearly regulated bulb will have less RF emissions, though you could theoretically still have a bit of rf trash from edge effects depending on the rectifier and filter cap.
Any bulb that has passed FCC Class B will have an upper bound on conducted and radiated emissions, but you can pass such tests on a wide range of margins, and therefore certification isn't useful to clarify which bulbs have the lowest emissions.
I think that we'll start to see improvements in this area as people move away from "screw-in bulb" as the form factor and towards using external power converters delivering power to to LED fixtures that are just LED's and their heatsink + mounts.
Perhaps USB-C PD might help? USB power adapters are compact and PD offers standard, useful high power DC. If the adapters become ubiquitous, you could see "USB-C lightbulbs".
- A shock from a DC supply will cause heart fibrillations at lower voltage than AC, because of the way muscles receive signals from nerves
- An electrical arc from a DC line will burn continuously, while an AC arc tends to self-extinguish at each zero-crossing.
So in practice, 120 VAC or 48 VDC are the limits for household wiring.
About 20% of the world uses 110-120 V and 60Hz. The rest uses mostly 240V at 50Hz.
Although 110V is safer, the US seems to ignore many of the safety features I see used elsewhere in the world.
How that translates to deaths per million people per year comparing to similarly rich countries, I don't know.
That is opposite to my understanding. The extra shock hazard of AC vs DC was an argument used against AC power distribution back in the day. The 50V limit is to prevent shock altogether.
If you had AC to your house, then a rectifier for a 100V/200V DC circuit through your house, the losses wouldn't be any worse.
DC losses are less than AC losses at the equivalent voltage, because the resistance seen by DC is less than the AC impedance.
In a perfect world we'd have both AC & DC, and could use whichever was most appropriate. We could feed DC straight from the solar panels to the cell phone. But in terms of what consumes most of the power, in the typical home it's AC loads.
This isn't an appropriate comparison. Those appliances actually do useful things that need to get done so the energy isn't just wasted like it is with power supply inefficiency.
Sure a 12vdc circuit would be nice. It wouldn't be nice enough or efficient enough to be worth paying for.
I hate waste; my mains-connected smoke detectors (1% efficient) and garage door opener (15W standby) gnaw at my soul. But that's scope for improvement, not an efficiency crisis.
Those aren’t a waste, they are critical life safety equipment. If their 1% efficient power supply lasts the full 10 year lifespan of a smoke detector rather than 6-18 months like a typical LED lightbulb, I’d say that’s the appropriate level of reliability.
If every home in Australia had two smoke alarms, that means 5,600 kW of continuous and largely wasted energy consumption. I don't mean wasted in purpose: smoke alarms are essential. The wasteful part is the fact that 99% of the power going into smoke alarms goes to converting AC power to DC power.
As far as I'm concerned, that's 99% scope for improvement.
https://reductionrevolution.com.au/blogs/news-reviews/584256...
Although 15W standby for a garage opener seems excessive.
Just because generator/machine set transformers and 380/220 kV step transformers require semi-active cooling doesn't mean they're inefficient... just means that they handle a huge amount of power (MWs), so even at very high efficiencies that translates to a lot of heat in absolute terms.
Sorry, I was being a little too loose for HN. Also, I'm not an EE but I play one on the Internet :).
You can buy PoE LED light fixtures now, which are both powered and controlled over ethernet. You buy a single PoE switch, run some low voltage Cat6, and you can power an entire floor's worth of lighting.
EEPoE (Energy Efficient PoE) offers more efficient PoE too which claims up to 94% efficiency.
Definitely a space to watch. Particularly as the cost savings of running low voltage Ethernet cable compared to high voltage electrical cable (110v) are substantial.
They tried every random thing, including putting a small real ARM server on the bloody thing to do remote control. It was back then when I first stumbled on Espressif people and Teo. They got lightyears ahead of us with all-in-one SoC.
Microsemi's exclusive EEPoE technology cuts the power losses on Ethernet cables by 50%, through the utilization of all the copper available on cable when a Microsemi EEPoE PSE IC or Midspan is used. It is 100% compatible with IEEE802.3at, and the savings work with ANY IEEE 802.3at Type 2, Type 1 or IEEE 802.3af compliant PD. In practice, devices that consume 25.5W would consume less than 27.75W, instead of the worst case 30W when a non-EEPoE PSE is employed.
So they use all 8 wires for power instead of just 4 to halve the power loss on the wire. For the worst case device that makes the 20% loss be a 10% loss.
? Must be either very low current lighting or that's some pretty beefy Cat6. I think the average Cat6 cable is only 24 gauge and that won't carry a lot of current very far.
Biggest problem tends to be accommodating the central power delivery device. Those big PoE switches run hot and loud.
I always imagined homes would migrate to 12VDC and converge electronics with RVs.
And it's not only hams, but also other radio spectrum users, like mariners -http://www.professionalmariner.com/December-January-2019/Onb...
These already exist. Nearly every architectural grade fixture has a separate 24VDC driver, the LEDs, and a heatsink. Here's a cutsheet for a Lithonia recessed can as an example:
https://lithonia.acuitybrands.com/products/detail/748533/Lit...
These run for $175-200 a piece. They list 70% lumen maintenance at 50,000 hours. That's 20 years at 12 hours a day for 220 days a year. Some architectural cans are a bit cheaper, some are a bit more.
Cheap LED 2x4s go for $80, pricey ones for $200.
I've been using them for a couple years, and think they are so much better than the previous generations of CFL's and heat sink LED's that would quickly fail, especially in enclosed fixtures. I'm buying cheap ones from Amazon, and occasionally I get early failures, but lifespan after initial burn in has been excellent. Like others, I often wonder why they haven't taken over the market: http://www.eevblog.com/forum/chat/why-haven_t-filament-led-b...
Because most POS systems now use a thermal printer system that for some reason - even when kept at a "proper temperature" - the receipt fades over time, to where it is virtually illegible.
The conspiracy part of me thinks this was done on purpose to reduce returns and warranty claims...
I had a Lowe’s receipt do exactly what you describe out in my shed.
i.e., if you agree to be tracked.
It's like GPS tracking: I track every trip I make with OsmAnd's tracking feature automatically when it starts routing. But I wouldn't trust Google Maps to do the same with their location history feature. Their usage of the data isn't limited to what's in my best interest, so I'll collect it myself.
This is relevant to my interests...
AIUI in EU we get default 2 years of warranty on all goods. And we get a 14 day cooling off period on goods that are the result of solicited sales (cold [phone-]calling, selling at your door), or are bought online.
If an LED bulb should last 10 years then up to the 6 year point you should be able to get 40% back on a failed unit. Up to 2 years, get a replacement, repair, or full refund (at the sellers choice).
With the CRA the right is against the /seller/ and not the manufacturer though.
AC circuit would still be used for things that have motors, but everything else uses DC.
Imagine how simple LED lighting could be on a DC circuit. Just an LED and resistor.
I only took a few electrical engineering classes so correct me if this is a bad idea.
High voltage DC is not useful for the things people use.
For one, you don't need a transformer for stepping down voltage, as you can see in essentially every piece of consumer electronics nowadays. Your typical CPU nowadays runs in the 1-2 volt range, but PC and laptop power supplies and batteries provide various voltages between 3.3 and 20 V. Your typical desktop mainboard takes most of its power to feed the CPU from the power supply's 12 V rail. Yet, you won't find a single power transformer outside the power supply--the only transformers you will find on a mainboard are probably Ethernet signal transformers.
All of that is done by switching regulators, and those work perfectly fine for regulating 120 V down to 5 V as well.
Then, rectifiers really don't have a problem rectifying 120 V. Some of the cheapest diodes you can find are perfectly capable of rectifying 120 V mains voltage. In fact that works so well that almost every power supply, at least for consumer electronics, nowadays does just that: The input is fed through a rectifier and a filtering capacitor to first to procude (in the case of 120 V input) ~ 170 V DC. After that comes some sort of oscillator that then generates more or less AC from that, but at a much higher frequency, in order to feed that through a transformer--primarily for isolation, but if you have to have a transformer anyway, you might as well use it for reducing the voltage as well. The output from the transformer then is rectified again for most applications, which is a lot more difficult to do efficiently due to the high frequency and the low voltage.
It would be nice if some people got together and wrote a standard for DC powered homes. Preferably without a bunch of patented stuff included in the standard. Then all the pieces could be produced by various industry players and we could switch.
I've built a lot of LED lighting setups, and even with the most efficient chips and converters, extremely underdriven, the footprint required to run efficiently and dissipate heat in a way that will let them last 10-20 years is an order of magnitude larger than what would fit in a bulb.
I've only done this in the living room as an experiment, and have no intention of doing this anywhere else exactly because of the maintenance overhead and the breaking with convention.
> But the fact that I'd have to sell my house one day stops me.
“Here. It works now. Here is a manual that describes how it works, here is a box of spare parts. You are welcome to rip it all out and replace it with conventional gear. There is no warrantee on this thing, even though it's awesome.”
The Hue app kinda sucks but those bulbs have been going strong for 4-5 years. No failures of course, and one bulb that was left outside (covered) in NYC and always on for almost two years. No issues at all, and runs very cool.
This is what amazes me, I've been an early adopter of LED bulbs (Google even gave out a bunch at one time to employees) and have seen my share of failures, from color shift, to brightness degradation, to controller failure. That said, they nearly always cost less to operate than the previous incandescent lamps they replaced.
When the LED 'tubes' which replace fluorescent bulbs became cost effective I replaced all the bulbs in my garage. And when a credible can light with acceptable dimming came out I replaced all of the recessed lighting in the family room and living room. I now don't have a single incandescent lamp in the house and it has reduced by monthly energy by anywhere from 35 to 55 KWhr per month. That represents about a $50 a year in energy cost savings. The cost has come down to the point where the bulbs pay for themselves well within their lifetime range (my experience over a population of 62 different bulbs has been 7.5 years MTTF).
That said, I haven't been as diligent in figuring out exactly why a bulb has died, if it is just the power supply I expect they could be resurrected less expensively than a new bulb, caveat the reassembly of the light diffuser bulb.
Not that using incandescent bulbs is a good option anymore anyway.
GE even had a warranty so I shipped them back on principle (shipping was nearly the cost of the bulbs themselves) and GE never responded with replacement bulbs. What a crock.
White lights emit more blue light than yellow lights, which can negatively affect your sleep. It's the real-world equivalent of uninstalling redshift/f.lux on your device.
I've invested in "smart" LEDs and I change their color based on the position of the sun. If the sun is up, the light is white. If the sun is down, the light is dimmed orange.
The first driverless, flicker free LEDs on the market were their invention. Basic idea is that if you fine tune phosphor components, and their ratios, you can make a direct 120hz AC driven LED with close to no visible flicker for as long as your output colour temperature is in between 5500 to 6000 Kelvins
The problem... they got copycated within months, even though copycats did not understand the importance of a fine tuned, custom made phosphor.
color temperature is really important. I despise 3300K+ headlights and bulbs.
Philips has a dimmable LED bulb where the temperature varies between 2200K and 2700K. I've also found standard 2200K bulbs. But mostly I prefer using single-color red orange and yellow bulbs.
My understanding is that white LEDs are basically blue/ultraviolet LEDs with phosphors to knock down the energy level of the photons. Do you have any links about the phosphor components used for different color outputs?
http://xlgy.scnyw.com/product/powder.html
http://en.scnyw.com/About/Affiliates/4.html http://xlgy.scnyw.com
They were bought and sold like 10 times over since I left them in 2013
Cree has a good 10 year warranty though, I've sent in boxes of dead bulbs to them. However many substitutes are dying out too and they're included in many of the replacements that are going back for my next batch.
Thus it seems that, just like with incandescent bulbs, LEDs will last practically forever if they're run dimly. The difference is that LEDs are actually more efficient at lower current densities, where they're converting more of their input power to light instead of heat, but cost considerations has lead to typical bulbs driving them at much higher currents for more brightness but correspondingly less life and efficiency. They're still far more efficient than incandescents, so I guess that's what counts...
but they are sort of illegal thanks to:
wikipedia.org/wiki/Energy_Independence_and_Security_Act_of_2007
"A light bulb at a fire station in Livermore, California, has been on almost continuously since 1901. In 2015 it was recognized by Guinness World Records as the world’s longest-burning bulb.... Giles Slade, in his book “Made to Break,” traces the term “planned obsolescence” to a 1932 pamphlet, circulated in New York, titled 'Ending the Depression through Planned Obsolescence'.... the lamps market will bring in an estimated thirty-eight billion dollars this year, L.E.D.-bulb makers are already reacting to the spectre of declining sales."
ESR limits current inrush into the capacitor, with ceramic caps you would need some additional current limiting.
Voltage should not be that much of a problem, but there may be other issues with ceramic caps (like cracking due to thermal stress, voltage dependent capacity, microphoning, ...)
Plugged it in, did not work. Tried dismantling it with the help of these photos — https://www.youtube.com/watch?v=jhfZ-vwau8c — but some screw sockets seemed damaged and the screws would not come off. I could not see anything wrong with cables or soldering as far as I got, so I wonder if the LED or circuit board is the problem.
Good design needs to consider sustainability, longevity and ease-of-maintenance in addition to how something works and looks.
Appliances back in the days could last decades. Now you're lucky if they don't break in a few years.
What do they say about the best type of customer? The best customer is a repeat customer :)
By the same token, all the large CFL and LED lamps in my house are expensive high-CRI lamps. I've seen two CFLs go due to a failure in internal electronics, likely as a result of a power glitch. The longest-running CFL I had to date ran for 7 years, for 12+ hours a day.
For most products I find this to simply not be the case anymore. What I find is the more you pay, the more "gizmos" and "features" you get that ultimately lead to earlier device failures than the cheap models.
A 1962 washing machine was 184.95 [1] ($1523.97 today) and a 1959 model was $209.95 [2] ($1729.97 today); which are both more expensive than all but the highest end front loader currently at Best Buy. The vast majority of top loaders are in the $500-700 range, top loaders $800-1000. There simply isn't much market for $2000 machines that will last decades.
It's a similar story for almost all home appliances I've looked at.
1) http://www.thepeoplehistory.com/60selectrical.html
2) http://www.aei.org/publication/appliance-shopping-1959-vs-20...
Those "high end" washing machines will also fail in the same time range that the low end ones will.
There is a market for them, but it will remain unfulfilled because the manufacturers make more money in the short term with the low-reliability ones.
I just discovered that a few days ago because I was using a LED maglight to test my panel and realized I had incredibly small power output with it. Now I am building a small circuit to measure and datalog the power output of different solar panels under different type of artificial lights (vs having to use the multimeter + oscilloscope which is not very portal outside my labs).
If you designed solar powered device for indoor use and have heard of this issue, I'm very interested.
Sure, eventually, but without further context this obsession with light bulbs always sounds a little bit penny wise and pound foolish to me.
The relative cost of lighting on the monthly electricity bill is tiny so those savings are pretty small. Getting a more efficient fridge, washing machine, TV or (in our profession) computer typically has a much bigger effect on electricity expenses. And that is assuming your heating and cooking is done with gas.
Thus environmental policy centers on shopping bags, straws, and light bulbs, rather than things that objectively have real environmental impact.
That said, leds that are 13-15 watts instead of 60 watts does make a big dent into saving electricity at the country scale.
A single bulb used 5h/day[0] on average means 90kWh/year. Aside from lowering the load somewhat, depending on the electricity prices the LED bulb might pay for itself in a year.
[0] which is not really difficult, 6 to 7 in the morning and 18 to 22 in the evening and you're there.
https://www.forbes.com/sites/williampentland/2011/05/05/bill...
If we're going to fight climate change we're all going to be forced to give up most meat anyhow, so might as well start.
The switch from using tiny space heaters for light to using LEDs that consume a tenth of the power was a huge savings. It only takes a few 100w bulbs to equal or exceed the energy consumption of a modest desktop computer.
Which countries have widespread recycle facilities for CFLs? In UK I'm almost certain they all just go in to landfill (in this country or another).
However yes, most people don't know about that and just landfill them.
Bit out of date but indicative of the UK situation I imagine.
The best I could do for the USA situation was that [1] shows about 1.5M recycled CFLs in 2010, vs [2] shows about 300M sold in 2009, 0.5% recycle rate ...
[0] https://conversation.which.co.uk/home-energy/do-you-recycle-... [1] https://www.consumerreports.org/cro/news/2011/04/despite-mer... [2] https://www.masslive.com/news/index.ssf/2011/12/cfls_compact...
I just thought that "everybody" has switched to CFLs 10 years ago, when they became decent.
LPT: Don't buy the globe ones that go over your bathroom mirror. They are practically pre-burned-out for your convenience.
Maybe you should. I mean, the cheapest option is rarely the most durable or performant. See Vimes' Boots Theory.
https://en.wikipedia.org/wiki/Sam_Vimes#Boots_theory_of_soci...
Might even take a single one, depending on the specific machines it might be pulling no more than 60~80W idle.
As in performance optimisation, energy optimisation requires measurements.
That's an order of magnitude more than an idle laptop.
And they're not talking about any sort of "average system" they're ballparking a "reference desktop" which tops out at 150W under load, that's not just "don't use a gaming GPU" that's "don't use a GPU at all". And if you're building a "desktop computer" with an APU and an SSD… why are you building a desktop computer?
Uuh.. yes... And an order of magnitude less than "several 100w light bulbs"
> And if you're building a "desktop computer" with an APU and an SSD… why are you building a desktop computer?
Ergonomics?
Only because of the successes up to this point. Marginal gains in LED efficiency aren't going to do much, but the shift from incandescent to LED has had a big impact.
I have LEDs in my house, but only because my local utility pays for most of them. If not for that I'd use only CFL's - they are cheaper, and just as good.
LEDs also don't care about cycles. A CFL doesn't like being cycled on and off very much. So that leaves a temptation to say well, I'll be back in here again in a minute and leave it switched on. No temptation to do that with LED.
Might be, might not be, it all depends on the usage of the various items e.g. let's say I've a 100W incandescent kitchen bulb, it's lit 5h/day on average, an equivalent LED would be 12~15W for ~$5 and will save ~3000Wh per week.
Going from a Class A to an A+++ washing machine would save ~500Wh/cycle. And I can buy 50~100 bulbs for the price of the washing machine, so that'd be both penny foolish and pound foolish.
Maybe "without further context" you should avoid hard and fast assertions, should run the numbers and should help people try and profile their energy usage instead of putting down their efforts?
The major energy cost for clothes washing is heating the water. A rooftop solar water heater dramatically reduces energy use, often more than one would get from using the same space to install solar panels.
A front loading washing machine uses less water than a top loading one and will also use less energy for water heating.
Most modern washing machine are cold-water only, washing machines with a hot water inlet tend to be prosumer. Same with dishwashers.
Wait, are you talking about the machines that have an electric heating element inside the washer? And you can still wash warm and hot cycle, but they don't require a connection to the hot water heater? Or are you talking about machines (which I've never seen but am willing to accept exist in other countries) that actually only offer cold cycle as a washing option?
If we're talking about ones with an internal heating element, those generally use about twice as much energy to operate than ones that rely on a hot water intake. But with the two input machines one has to figure out the energy to heat the water as a separate cost, which is what I was mentioning. Gas heated hot water heater will be more efficient than electric heated, and solar heated will be most efficient of all.
The former. I don't think I've seen a consumer-grade dual-intake washing machine in 10+ years.
OTOH Japan might actually have the second, I dimly recall my sister mentioning something along those lines.
And after I've done all of that? Or maybe I'm looking for easier, less expensive wins than dropping $700 on a fridge? Let's not poo-poo the choice of replacing inefficient, CO2-producing tech by being quick with the "yeah, but...".
Or, switch entirely to Philips/Ikea smart bulbs which are very affordable now. I think dimmers as a wall-mounted fixture will be on the way out soon.
In the UK, for example, during winter (and spring and autumn) the heating is appreciated! However, it is 2-3x more costly than using gas central heating. (However however something something whole house vs room-specific heating.)
In the summer, it’s a complete waste though.
Directed infrared heating panels may be the most efficient because they do what you describe - heat you up directly, rather than heat the air in the room.
It's a straightforward issue of separation of concerns, there are some occasions where you want heating and lighting always tied together, but they are niche. Almost everywhere in the world you have large periods of the day or year when you want lighting without heat.
https://en.wikipedia.org/wiki/Heat_pump#Performance_consider...
It separates the light and heat concerns.
Could only find one video of the effect: https://www.youtube.com/watch?v=wheumkzDslA
Could it be that with the initial 100k LEDs the definition was different than with current ones? As mentioned, the industry is now using the 70% of brightness as the limit. Since such figure was probably not defined early on, maybe the vendors put the limit much lower.
The CFLs I bought in the mid 90s lasted for years and years. Now they are practically pre-burned-out for your convenience.
The article goes into how the color and brightness degrades naturally over time, so it may also make sense that they counting more of that behavior as failures rather than the traditional "burn out".
> Locate the Weakest Link: Component Lifetime
This is why I tell people to buy simple "driverless" LED bulbs. A shitty constant current driver is worse than not having it at all.
I'm curious how it compares to the LED filament lighting which can use much simpler power supplies with fewer components which should mean much lower failure rates.
Edit: the source is not mentioned anywhere. This is an excerpt from Gravity's Rainbow.
Eventually I'll replace the rest of them as they fail with whatever cheap store brand LED bulbs I can get.
The fittings are open-backed (not ceiling-recessed) which probably helps.
In pursuit of cost savings, we made a simple product 100x more complex.
Seems like some municipalities got sold a pile of crap. At least they work even with the missing lights, but who knows how often they are having to replace those?
All the failures I've torn down over the years had failed power supplies and perfectly functional LED's. I've got a drawer full of the for little projects.
I might go back, I do not like the look of my house, and I find myself turning on my living room lights AND dining room lights. The brightness wasn't there.
There is massive improvement to be made on LEDs still.
CRI, lumen, color temperature and watt (actual, not incandescent-equivalent) are the relevant parameters. The item titles are generally vague.
Specifically for the full spectrum you want CRI > 0.95
G9 is probably quite challenging for high-performance LEDs because the size is so small. In the other sockets like E27 and GU10 there are lot of options available and also products designed for commercial use (where you need certain characteristics because you want to make your products look good under the light).
There's a great documentary on planned obsolescence for those that haven't seen it.