Longest lived standard fixture (retrofit) LED light bulbs?

[james_s]

I always buy the smalles package of lamp I can find first and spend some time evaluating the lamp before I buy more.

[cdev]

Wow..

https://spectrum.ieee.org/tech-history/dawn-of-electronics/the-great-lightbulb-conspiracy

"There were continual reports of cartel members’ attempts to restore the burning time of their bulbs to the old levels in defiance of the watchful eyes of Phoebus. At one point, some members surreptitiously introduced longer-lived bulbs by designing them to run at a voltage higher than the standard line voltage. After the Phoebus development department’s customary report of voltage statistics revealed such product “enhancements,” Anton Philips, head of Philips, complained to an executive at International General Electric:

“This, you will agree with me, is a very dangerous practice and is having a most detrimental influence on the total turnover of the Phoebus Parties…. After the very strenuous efforts we made to emerge from a period of long life lamps, it is of the greatest importance that we do not sink back into the same mire by paying no attention to voltages and supplying lamps that will have a very prolonged life.”

As this episode reveals, tweaking a lightbulb’s rated voltage was one way to modify the product’s life. Another was to adjust the current, as GE engineers did to decrease the life span of its flashlight bulbs. A GE flashlight bulb in the precartel days was designed to last longer than three changes of batteries. This life span was then cut to two battery changes, and in 1932 the GE engineering department proposed that the bulb last no longer than one battery. A GE engineer named Prideaux wrote in a memo, “We would suggest increasing Mazda lamp No. 10 from .27 ampere to .30; and 13.14 and 31 from .30 to .35. This would result in increases of candlepower of 11 and 16 percent respectively.” That boost in illumination, he suggested, “would be acceptable to all flashlight users” despite the fact that the higher current would shorten not just the bulb’s life but also the battery’s."

[DougSpindler]

Phoebus cartel, anyone?

https://en.wikipedia.org/wiki/Phoebus_cartel

Now seriously, I don't think there is a similar cartel for LED lamps. Rather, as one has explained, the culprits are the sh!itty drive electronics and/or poor thermal management.

Like James_s, I also write the installation date on the body. I have a few installed outside with an automated photocell turn on/off, and so far have worked reliably every night for close to 8 years. They were made by Toshiba.

Also.........My living room ceiling is tall (18 feet) with several ceiling-recessed lights that would be an incredible pain in the anode to replace.
In the long and sweltering summers down here, even with the extra insulation, the ceiling on its own becomes quite warm. 

I know that heat is an effective electronics killer, therefore for those lights I wired a dimmer which is permanently set at 80% level. (The bulbs are dimmable), in an effort to run them cooler.
I don't have enough data to claim that my theory will provide a longer life, but I strongly suspect it does.

You had better believe there is a cartel....  The Chinese cartel.

I doubt dimming will reduce the heat from the LEDs.

[james_s]

They run far cooler when dimmed. Why would you think otherwise?

[DougSpindler]

They run far cooler when dimmed. Why would you think otherwise?

Running at reduced voltage would increase the amperage resulting in less efficiency and same or more heat if were not dimmed.

[beanflying]

Amazing several times here flickering LEDS are the problem. Seriously on an EE forum flickering leds are not the problem the problem is the drivers and the supply! And to ecological damage of LEDs yeah because the Mercury in CFL's were awesome for the environment 

There is a similar thread running on filament bulbs here for those that haven't seen it. https://www.eevblog.com/forum/chat/fialment-style-led-bulbs/msg2027542/#msg2027542 'Some' bulbs are starting to run decent drivers in them but keep demanding $0.02 less cost on an item and you will get bad flicker. The IC on the photo below can be sourced for between $0.06 and $0.10 each here https://lcsc.com/product-detail/LED-Drivers_SM2082ED_C80717.html

Blame the drivers on the shitty end of the spectrum that YOU and the dumb ass buyers for the companies in YOUR countries buy not the LED.

Failure is a different issue and Overdriving and poor cooling is the principal cause and this comes down to wanting more lumens than the next guy for pennies saved be that less LEDs or less aluminium.

Bottom line is we need to start demanding better of the entire supply chain and pay a few more cents or we will continue to get flickering lamps.

[cdev]

I think you may be confusing LEDs with incandescent bulbs. Less voltage = less current with LEDs.

They run far cooler when dimmed. Why would you think otherwise?

Running at reduced voltage would increase the amperage resulting in less efficiency and same or more heat if were not dimmed.

[james_s]

And less current results in dramatic efficiency gains. If you look at the lumens per watt of a typical power LED, the efficiency drops pretty significantly driving them at 700mA vs 300mA.

I haven't measured but the LED bulbs I have on dimmers run at a temperature I can hold my finger on for a second or so before it hurts so probably around 60C. Dimmed to approximately half brightness they feel room temperature to maybe very slightly warm to the touch.

[cdev]

I have a strip of bright blue-green "piranha" I think the name is, LEDs not heatsinked at all, but very bright. They are not running at their rated current or voltage drop at all, though and are cool to the touch. They have been running continuously for at least 15 years, (with a break to make a cross country trip) on the same DC power supply with NO discernible drop in brightness. Not even the tiniest bit. I use them as a hall light but the current draw is so low that it makes no sense to turn them on and off, they are just on all the time.

[DougSpindler]

My flawed thinking......  Thank you.   I will stand corrected.

I think you may be confusing LEDs with incandescent bulbs. Less voltage = less current with LEDs.

They run far cooler when dimmed. Why would you think otherwise?

Running at reduced voltage would increase the amperage resulting in less efficiency and same or more heat if were not dimmed.

[IanMacdonald]

Blame the drivers on the shitty end of the spectrum that YOU and the dumb ass buyers for the companies in YOUR countries buy not the LED.

So the answer to any problem is to throw money at it.   

I think I need some $100-a-foot speaker cable. The cheaper stuff obviously must be shitty. 

Seriously, the problem here is that there is little incentive for manufacturers to fix flicker because it's not something that can be determined at point of sale, therefore they will see it as being unimportant  in terms of affecting sales volumes. I'm generally opposed to bureaucracy and red tape, but this is precisely the sort of situation where product testing and certification is valuable. 

[beanflying]

So the answer to any problem is to throw money at it.   

I think I need some $100-a-foot speaker cable. The cheaper stuff obviously must be shitty. 

Seriously, the problem here is that there is little incentive for manufacturers to fix flicker because it's not something that can be determined at point of sale, therefore they will see it as being unimportant  in terms of affecting sales volumes. I'm generally opposed to bureaucracy and red tape, but this is precisely the sort of situation where product testing and certification is valuable. 

And you are being a wally for what reason exactly?

There is and will be a minor cost to fix flicker so your point is that it should happen for free to the consumer because you think it should be so?

We are talking throwing a matter of cents per lamp here and the will to change and spend it is virtually zero because of the supply chain keeps pushing the low bidder model to maximise $. Pressure has to come from somewhere a point we seem to agree on. And we agree on it being a good idea.

Certification, Regulation and then consumer education on a world scale lets get the EU to do it they love this stuff then adopt it worldwide as this certainly WON'T be FREE to implement.

Unless in your mind you have another way to make that free too?

[cdev]

If people were just a little more literate electronically, they would realize that better heat sink + slow fade out when power is removed = almost certain to be a better LED light.

[DougSpindler]

But that would requires knowledge and education.  That’s something businesses don’t like to spend money on.

If people were just a little more literate electronically, they would realize that better heat sink + slow fade out when power is removed = almost certain to be a better LED light.

[james_s]

Very few of my LED lamps do a slow fade when you shut them off, and they have been long lived and don't flicker. I suspect it has to do with the way the driver handles dimming.

[schmitt trigger]

What I meant above of a dimmer running at 80%, it was at 80% power, not at 80% voltage. Measured with a Voltech power meter.

While I was doing that, I was simultaneously measuring the lux level at a center table. With 80% input power, I was only reducing the luxes to 87% of the full power level. A nice power consumption reduction with a barely  discernible reduction in illumination. This corroborates James's statement that the efficiency decreases at the highest power levels. Come to think of it, this fact on itself would reduce heat generation.

And yes I understand the difference between lumens and luxes.

[james_s]

Looking at a few datasheets, it seems the light output is a lot more linear with increasing current as it was in previous generations I had worked with. The output does still drop with temperature though, at 100C it is around 18% lower than it is at room temperature on the Luxeon medium power LED I just looked at.

[electromotive]

I've had good luck with Phillips / Osram. Even the once mighty GE bulbs have failed me. I'm wondering if the GE bulbs themselves are just a licensing agreement, because they don't seem to last much longer than the no-name subsidized bulbs here in the states.

[DougSpindler]

GE for most consumer goods is a licensing arrangement.  Other than medical equipment and nuclear reactors I don’t think GE has made any consumer products in decades.

[kaz911]

For retrofit halogen I always put in Philips Master series. They are designed to have a “larger” turn on power consumption and “anti-flicker” so it usually works with old transformers. They might use slightly more power than other LED’s but have high CRI and rarely fail.

I have a lot of halogen that has been replaced - and I think out of about 90 bulbs where 50 are Philips Master - only 1 has failed in 3 years. But it was promptly replaced FOC by Philips.

For screw in fixtures I use Ikea’s latest - they have become pretty good for the price. The old ones had horrible CRI’s but the latest seems to be a lot better.

[ahbushnell]

The problem is LED bulb manufacturers are full of patent litigation. Over heatsinking, power supply, adhesives, phosphors, LED's etc.
It's so bad the past 10 years nobody can make a reliable, low cost LED bulb without going through hoops to not infringe on someone else's patents. So the bulbs are not the best designs, many compromises.

And now I see LED manufacturers suing the retailers now, Lowe's getting sued by Epistar, Nichia suing Bed Bath and Beyond or Feit Electric, LG, Lowe's, Meridian, Everlight etc.

Ikea store lighting section and see a handful of LED fixtures flickering and cutting in and out.

Cree LED bulbs, I had about 25% of them fail within a year. You have to ship them back to Cree at your expense to claim warranty. So I ditched Cree and Home Depot also dumped them.

I buy LED fixtures for the garage.  The LED's are spread out over a large area and the thermal load is low.  So far so good.  Good idea to record the date.  I never think about that.