| Electronics > Repair |
| Why/how do these LED lamps fail so fast? |
| << < (4/9) > >> |
| IanB:
--- Quote from: alligatorblues on September 21, 2023, 01:21:47 am ---If LEDs are run off constant current, the voltage remains steady. But if the current is not well controlled, the voltage will rise exponentially in comparison to the rise in current. It sounds like a poor design. The current creeps up. I doubt if it's worth trying to salvage. --- End quote --- Actually, this is not entirely true. Constant current supplies like to kill loads connected to them, and they are more deadly when the current is controlled more tightly. Here's an experiment I did: I took a length of nichrome resistance wire and connected it to a variable power supply with voltage regulation ("constant voltage"). I was able to gradually turn up the voltage until the wire glowed red, orange, yellow, quite brightly. Everything behaved fine. Then I took the same length of wire and switched the power supply to current regulation ("constant current"). This time, when I gradually turned up the current the wire glowed red, started to be orange, and then quickly burned out and failed with a molten hotspot. I tried again, two more times, with a fresh piece of wire each time. Each time the wire rapidly melted and failed before getting much beyond a dull red heat. This is the problem with constant current supplies. They are really unforgiving and like to destroy the things they are powering. |
| T3sl4co1l:
--- Quote from: IanB on September 21, 2023, 01:37:17 am --- --- Quote from: alligatorblues on September 21, 2023, 01:21:47 am ---If LEDs are run off constant current, the voltage remains steady. But if the current is not well controlled, the voltage will rise exponentially in comparison to the rise in current. It sounds like a poor design. The current creeps up. I doubt if it's worth trying to salvage. --- End quote --- Actually, this is not entirely true. Constant current supplies like to kill loads connected to them, and they are more deadly when the current is controlled more tightly. Here's an experiment I did: I took a length of nichrome resistance wire and connected it to a variable power supply with voltage regulation ("constant voltage"). I was able to gradually turn up the voltage until the wire glowed red, orange, yellow, quite brightly. Everything behaved fine. Then I took the same length of wire and switched the power supply to current regulation ("constant current"). This time, when I gradually turned up the current the wire glowed red, started to be orange, and then quickly burned out and failed with a molten hotspot. I tried again, two more times, with a fresh piece of wire each time. Each time the wire rapidly melted and failed before getting much beyond a dull red heat. This is the problem with constant current supplies. They are really unforgiving and like to destroy the things they are powering. --- End quote --- Actually, this is not entirely true. Constant voltage supplies love to kill loads connected to them, and they are more deadly when the voltage is controlled more tightly. Here's an experiment I did: I took a string of LEDs and connected it to a variable power supply with current regulation ("constant current"). I was able to gradually turn up the current until the LEDs glowed painfully bright. They got warm and that was about it. Then I took the same string of LEDs and switched the power supply to voltage regulation ("constant voltage"). This time, when I gradually turned up the voltage, the LEDs glowed dimly at first, but as they started to heat up, they got brighter and brighter, until suddenly eclipsing the previous test and then nothing. I tested the LEDs and found one had failed open. I tried again, two more times, with parallel arrays of rectifier diodes this time, hoping that their robust ratings would do better. The LEDs were only rated a few hundred mA after all. Alas, each time, one of the diodes failed shorted, while the rest hardly even warmed up (WTF!). This is the problem with constant voltage supplies. They are really unforgiving and like to destroy the things they are powering. :) Tim |
| T3sl4co1l:
Anyway, regarding LEDs: they are extremely long lived, in and of themselves, when ratings are respected. The phosphor is actually the higher wear element, as it tends to fade over time. The power density on it is quite impressive, for just a powdered mixture of ceramics that have this one magical property (fluorescence). The blue LED chip itself also declines, but slower. But that's just the LEDs themselves. What's powering them, is another matter. Lifetime can always be lower, and when it is economical to do so, it absolutely will be done. Electrolytic capacitors are the most obvious and worst offender, with 2khr 85°C parts being common picks. Put that in a stuffy lamp fixture and it might last mere months! To make a long life LED lamp, one must use: - Quality LEDs, derated modestly to reduce phosphor and LED chip wear - A robust controller/driver, to limit current, typically to smooth light output, and optionally to perform power factor correction - Long life or low-wear components (e.g. 5khr+ 105°C electrolytics, if any) - Adequate transient protection and filtering (suitably sized MOV, somewhat-thicker-metallized EMI caps) - VERY GOOD thermal management, spreading heat from the LEDs throughout the package, and allowing air to permeate it - This includes the fixture, which should preferably be an open vertical or horizontal design, or made with holes or louvers in the reflectors/diffusers to allow free air flow. For pretty much everything, heat is the killer. Phosphors, despite being mere ceramics, are indeed sensitive to temperature, and seriously degraded when ran hot. LED chips accumulate defects faster when hot, and when ran at high current density, as well. Electrolytics notoriously degrade quickly when hot. So do other capacitors, and other components generally. Resistors, film capacitors, and semiconductors (other than LEDs), generally are fine within ratings of course, and those listed above are the major wear elements. Filtering and transient protection are key to long life through environmental disturbances, especially surge which will progressively (or outright) damage passive / unfiltered / bare LED strings (and most likely filtered ones too, since the filter capacitor will only be 10s of uF, not enough -- nor nearly low enough ESR -- to swamp the pulse). LED derating also has the knock-on benefit of lowered resistive losses (higher efficiency). On top of this, for a variety of reasons, modern LEDs are closer to 3.0V drop at typical operating levels -- they're very efficient. Compare to the oldest (~first generation illumination) which were typically 3.6V or more, and dissipated more like 90% of power input (which was still an impressive showing, as illumination technology goes). The optical (quantum) efficiency is remarkable: modern LEDs produce so much light that the power dissipation is, not just palpably, but substantially, lower than the power input. Compare a first-generation 100lm LED (might be an 8mm dia. package rated 3W), running near ratings, to a modern one at the same intensity (which might be a 3x3mm DFN, rated for, actually about 3W still, PCB permitting), dissipating under 1W! Now, of a marketed product, obviously you can only establish so many of these things, even as an engineer willing to (potentially) destroy a few lamps in the process of evaluating them. As a consumer, you only have whatever claims the manufacturer makes, and 3rd-party testing. Some of the above points can still be evaluated, at least to a rough degree, as well as other points made in the posts above. As far as temperature, besides fixture selection (that's your own responsibility as decorator!), large size (more dissipating area) and more open structure (same!) for a given power rating is a potential indicator of performance. Also, regarding Philips specifically -- they made a remarkable early lamp, with I think still quite impressive lifetime today (though I think it's long since discontinued, I haven't checked?). The trick was making a large, replaceable phosphor module to place on top of the LED chips -- thus running cool, giving diffuse emission, and addressing one of the primary concerns at the time. Since then, phosphor lifetime has improved, as well as overall efficiency, and overall cost most dramatically -- this was back when LEDs were very new and rather pricey, IIRC, so the maintenance option had some promise. Whether that was an acceptable compromise for consumers, or just a good trick to convince them to buy the things, I don't know, but, I guess it's funny that it was a losing proposition in the long run -- in that, LEDs are now so pervasive and cheap that such steps are irrelevant, the whole bulb is easily disposable. (Which, let me see here. Here's a press release about them: https://www.ledsmagazine.com/company-newsfeed/article/16691970/philips-unveils-worlds-first-led-replacement-for-most-common-household-light-bulb Teardown: https://youtu.be/b4ZBfmLGRrk MSRP about $50: https://www.led-resource.com/2012/09/philips-l-prize-award-winning-led-bulb-review/ I don't see an article about when they were discontinued, maybe it only lasted a few years even. Well made unit for what it is, just not very competitive!) You can also choose just completely different designs that literally never existed before. Well, sorta kinda. Fluorescent tubes are linear sources, obviously, and glow discharge systems have existed before, but tubes are rather rigid, and glow discharge requires custom glassware. LED strips can be placed anywhere, easily, and trimmed to length (with minor restrictions). They do need to be wired up correctly, but once installed, they are distributed sources, that run cool, while giving diffuse wide illumination. Combined with color customization options, whole new design/decoration spaces are possible. |
| soldar:
--- Quote from: IanB on September 20, 2023, 11:42:02 pm ---This seems to depend on the brand. I am getting really good results with Philips bulbs and a Lutron trailing edge dimmer (the Philips bulbs being the most important part of the equation). --- End quote --- Thanks for that link. The video is very interesting and informative and I will keep that information in mind. The thing is that I am replacing ceiling fixtures, not lightbulbs. I will have to see if Philips makes similar fixtures. |
| themadhippy:
--- Quote ---The thing is that I am replacing ceiling fixtures --- End quote --- --- Quote ---And what ticks me off is that these are complete ceiling fixtures where you can't just change the light-bulb like in the good old days. --- End quote --- why are you repeating your mistake, replace the fittings with something that has a standard lamp base |
| Navigation |
| Message Index |
| Next page |
| Previous page |