Not quite.
We were fairly early adopters in the move to LED bulbs, replacing the four 60W incandescent bulbs I'm sitting under with LED bulbs in October of 2012. One of those LED bulbs died around a month ago, and another one failed last week. Naturally, I didn't have a Watt-Hour meter on them, but they lasted on the order of 1600 days (thanks to this site). How many hours per day? It's hard to estimate, but I feel comfortable saying 12 hours/day or about 19,000 hours. That's probably a worst case estimate. At the time, 4-1/2 years ago, people were saying LED bulbs were likely to give 50 to 100,000 hours. 50000 hours is over 11 years, if the bulb ran 12 hours a day. Clearly, something is amiss.
Within a few days of the overhead bulb failing, I had one in our kitchen fail. The one in the kitchen was bought a couple of years later than those first bulbs and was chosen because it's a Cree and I know from my career that Cree is a semiconductor maker that knows what they're doing. Unfortunately, I don't know exactly when I installed it, but it had a date code on it that indicated the bulb was made in late 2013. I believe that bulb could have been on 14 hours/day; there's a big range here, and that would be closer to the worst case. 14 hours/day since January of 2014 is still less than 17,000 hours. A visit to Cree revealed that the bulb should have been guaranteed for 25,000 hours.
What's going on here? I have yet to find an LED bulb that makes 25,000 hours, and almost certainly not even 20,000 hours. Unfortunately, I didn't do the one thing that would really help me be certain and write the date the bulb was installed on the base or side of every LED bulb. So I wrote Cree's customer service email address; I listed their part number for the bulb (their photo above) and said I wasn't sure of my numbers, but believed it didn't have more than 17,000 hours. Then I asked what I could do to get the life they're claiming. I didn't say I wanted anything other than information. I was hoping to get a note telling me how to get the most life out of the bulb - whether that meant I should direct chilled 30 degree Fahrenheit air at it whenever it's on or I need to sacrifice the blood of a virgin chicken. So I was surprised to get this email the next day:
Thank you for taking the time to contact us. We apologize that your experience with your Cree LED bulbs was unsatisfactory. A warranty request has been processed and one replacement bulb will be shipping on 4/18/17 via FedEx or UPS. It is not necessary to return the defective bulb; it may be disposed of, or recycled.I replaced the burned out bulb with another 60W LED until the replacement became available. It's in place and quite good.
Our kitchen was redone in '03, and (as was popular at the time) setup with recessed halogen bulbs. Halogens are a nice light quality, but I had always heard they were more efficient than the conventional bulbs. About the same time as the LED failed, one of the halogen bulbs died. I noticed when I was putting it in place that the box only claimed 500 lumens out of this 39 W bulb. I had looked for an LED replacement for the halogen bulbs before and never found them. This time I told myself, the other bulb I just replaced was 800 lumens, why can't they make a 500 lumen LED bulb to replace halogen bulbs? They can. This time I searched for the right shape factor (PAR20 - a parabolic reflector bulb with 2 inch front), and the slightly rounded top version of that, the BR20. I found this bulb which is actually 675 lumens in the same package, so brighter than the halogens - at 10W input opposed to the halogen's 39W. I ordered three and waited for a halogen to die, figuring I'd replace them one at at time (there are six in use) as an experiment. It didn't take long for a bulb to die, just a couple of days, so we replaced it.
But I think I'm not Done done. I neglected to write the installation date on the bases of these new LED bulbs, too.
This comment has been removed by a blog administrator.
ReplyDeleteWhen we were forced to buy CFL bulbs, I started dating each on on installation. Imagine my surprise when I'm replacing them at similar intervals to the incandescent bulbs they replaced. Clearly not getting close to the promised lifespan....and don't they contain Mercury? Oh yeah.....thanks fedgov, wait, what's that? Ethanol in gasoline will be better for the environment too? 10% reduction in gas mileage and gummed up fuel system says otherwise, nevermind new power tools needed. Rant over.
ReplyDeleteThanks for the LED info.....time to buy more bulbs.
The only incandescent bulbs I own are in a box in the attic, my halogens are clamp-on 500 watt floodlights, and the only remaining fluorescents are the 8-footers on the garage ceiling.
ReplyDeleteHaven't lost any LEDs (yet), but they're not on much here at Castle Insanity; there's one 6W Cree on a timer in a UPS-managed floor lamp (for dusk-to-dawn target identification purposes) that's been on 12-14 hrs/day for 4 years, and an 18W Cree on 5-7 hrs/day for about 2 years. Lots of recessed lights, all 3-year-old LED retrofits, none used >3 hrs/day.
I wonder if your LED failure rate might be related to voltage variance. 3 years ago our Toy Electric Company killed the fridge by remote control; we used to get the regular 9 AM Saturday morning power outage (seriously, you could almost set your watch by it) in which volts went from the prescribed approx 120 to zero for 1-2 seconds, but once we got only double digits for about 8-10 seconds (estimated by the dimness of the last incandescent flood over the kitchen sink), then zero, then something much >120 for 3-5 seconds (based on the incandescent brightness), zero for a couple seconds, then back to the usual approx 120.
The display on the fridge was dark after all this, the control board disagreeing substantially with the voltage fluctuations. Fortunately, it failed in "full cold mode" rather than "off" so it wasn't a crisis while the dealer determined there was not a replacement board anywhere in the country. We moved stuff around to keep it from freezing during the 4 days it took for the dealer to deliver a replacement fridge under the extended warranty.
This got me to do three things: first, a vow to never purchase non-US made appliances because certain overseas manufacturers don't understand long term parts availability, second, put 1500VA UPSes between the refrigerator and the power company on both the refrigerator and chest freezer (the TV and stereo receiver already had one), and, third, examine what we were getting on voltages (I'd rather see a $130 UPS die than a $2300 refrigerator and the food in it; the freezer was much less but there's >$1K food in it. That the UPS will keep the fridge and freezer running for long enough to get the generator running is a plus).
What I found over a few months was a rather disturbing voltage variance - 129 as a high, 86 the low (actually, zero was the low but ithat came during the power outages). It seems most appliances can tolerate - and tolerate is the correct term – highs of about 126-132 volts and lows around 86-88; sometimes during summer "brownouts" voltages can briefly drop into the 90s, but lower voltages do heavily stress them, especially ones with motors. It's not uncommon to see a momentary 10V or greater voltage drop when a heat pump or central air conditioner starts (that's called 'locked rotor current" because a zero RPM electric motor draws maximum current at startup - that's why the lights "blink". I didn't have an easy way to record hertz, probably should have).
Eventually, enough complaints to the state's Board of Utility Regulation got some changes (my alarm system logs all events, including power drops, so I was able to supply a spreadsheet showing 38 zero-voltage events over a 2 year period, which doesn't include low voltage events above 96, the alarm system's "backup battery not charging" reporting threshold) so it seems our worst problems may be over; after 6 months of reliable 117-122 voltages I stopped recording them.
But I wonder what your house voltages are, and what Cree has estabished in its design standards as minimum/maximum "tolerable" voltages, and how much voltage fluctuations affect service life. I know just a little bit of excess voltage goes a long way to shortening integrated circuit life.
ReplyDeleteSince both of these comments go together, this is really a reply to both. In our experiences with the curly CF bulbs, we found what you did, Differ, only worse. Not only did they not last much - if any - longer than an incandescent, they usually failed with the power supply in the base emitting that familiar odor of burning electronics. Some of them cracked at the base, leaving the bulb portion hanging by wires, although none cracked and dumped their "toxic waste" in the house. None of them actually caught fire, in the open flame sense (thankfully), but the short life span of those bulbs got mentioned in trade publications (IEEE Spectrum, IIRC, and more). It was reported that bulbs mounted base up (hanging from a fixture) rather than base down (like a table lamp) failed more often.
The consensus was that there were two main problems. Part of it was that the bulbs were rated unrealistically, like regular fluorescents: turned on and left on until they burned out, not cycled every few hours like most people use their lights. The real heart of the problem was their design. Those bulbs have a small switching power supply in the base and they were really crappy designs. Normal, sane engineers derate their components; if you put in a resistor that's dissipating just under a half watt, you don't use a half watt resistor. You use the next bigger size. You can debate exactly what derating margin you use, but you always have margin.
This is where Nosmo's point comes in. The LED bulbs are also based on a switching power supply, but since the load of the LED is "easier" than the CFL tube, they're simpler (LEDs are more resistive so easier to design for). Still, those supplies have to be designed for a range of input voltages and if your supply goes outside of that range, damage is possible. Especially if the design isn't adequately derated. It's possible inadequate derating is the cause of both kind of bulb not lasting as long as we'd like.
My AC power tends to run high - or what I consider high. More like 121/122 than 115/117. While it's almost possible to grab a digital voltmeter in every room of the house, I don't have one plugged in the outlets all day, nor do I have a recording DVM, which is what I'd really need. I have a UPS on these computers and the TV, but not appliances, and I'd need protection on every outlet and E26 socket in the house.
On the other other hand, I might have a way to monitor AC with one of these doorstop computers I have lying around, or even my Raspberry Pi. It's worth thinking about.
I live in the boonies on a rural electric cooperative (supposedly cheaper than the other local provider - not that we have a choice - but our bill is $42 before we even use a watt of power), ands we have frequent power outages. Not just during winter (usually three or four that last several hours), but at least three or four during the rest of the year as well. I've got a 3.2 kw solar system with 830AH of battery back-up ($11k of 2v industrial batteries), but it is manually switched, not auto. The system is not set up to monitor incoming AC voltage, but I have never experienced what appeared to be a voltage drop per se, just periods where the power has gone completely off.
ReplyDeleteIs a quick drop to zero voltage as damaging as running at reduced voltage for a period of time? From the post and the comments, it doesn't sound like it, and my guess would be "no", but my understanding of electronics consists of training in the Navy (back in '69 ;-) of which circuit boards to pull on a failed radar unit, with a little old-fashioned transistor theory thrown in for laughs.
Do most decent UPS's also contain surge suppression? I _think_ I understand that operating at reduced voltage can be as (more?) damaging as slightly high voltage. I'm not worried about the solar system (Outback, not Xantrex), as it's output through the battery back should remain steady while it is running (and my generators are set tocharge the batteries, not supply loads directly), but fairly common outages and the occasional thunderstorm.
P.S. I _do_ know that my grid voltage runs around 121v AC here, not 115v. I have checked it often enough to be certain about that. Also, almost every bulb in our house is LED. I have lost a few LED bulbs, but they were older designs I've had for three or four years - made in China, IIRC.
DeleteSame here with the CFL's. They just don't last anywhere near as long as the claimed life, regardless of whether we use them base up or base down.
ReplyDeleteThe LED bulbs I've been buying at Lowe's or Home Depot have seemed to give good life so far. At least I haven't had to replace any yet, and the ones in use are about 3~4 years old.
I always label stuff like that with a Sharpie so I'll have some idea of their life span.
I'm still using incandescents. Fokk 'em for what they have done to this country.
ReplyDeleteI bought a bunch of incandescents when they were being discontinued, because I use them as heat sources for chickens and the shelter for an outdoor cat during the winter. They burn out fairly quickly, though, and I only have a few left. I'd buy more if I could. Most halogens and LEDs don't put out enough heat for my purposes, and most actual heat lamps use too much wattage.
DeleteI originally used CFL's in the barn as they ran cooler than incandescent and even used a couple in the house in globe type lights as they ran cooler and wouldn't burn out right away. That changed the first time we had lightning hit the wires during a storm and it took out every single CFL in the house and barn. Never bought another CFL after that. We have since used LED's and you just can't beat them. In globe lights and desk lamps they last much better than incandescent or CFL's and the best thing is they run cool which makes them much better in the barn and shop as well. I have far less of a "fire" phobia than in the past. They have also been getting cheaper and they use much less power as we run with lights left on in a few applications. So far it seems that the occasional voltage flickering in our rural location hasn't been a problem for LED's nor has total power loss from passing storms. indyjonesouthere
ReplyDeleteSo far the only house bulb LED failure has been a track light head and the home supply store replaced the head with no questions asked. (think orange shelving)
ReplyDeleteBecause of the high failure rate of the "new and improved" florescent tubes we have begun a replacement program of our four foot fixtures with admittedly more expensive LED fixtures. I did notice that the same seller that replaced the LED head is now seller LED replacement tubes that are plug and play and a little more research is needed.
I have three spare T5 fluorescent tubes left. Just replaced one about two weeks ago. When those three are used up, no more for me. I'm going to look into those LED tubes that are AC/DC to replace them with.
ReplyDeleteAn LED fixture should be cheaper than a regular fluorescent fixture: it doesn't need the ballast transformer. If you've ever had the ballast in your fixture die, you'll know that replacements are more than the entire "shop light" package you can get at that orange Borg.
I bought two LED T4-style lights for my reloading bench. They work great, but were fairly expensive, due to market economics. When they become more popular, the price will certainly come down.
DeleteHere is an article for replacing shop lights with cheap LED light strips, and they actually put out quite a bit of light, mounted on old shop light fixtures with the tubes removed. Worth a look if you haven't already seen this or something similar. I bought some, but haven't gotten around to doing the project yet:
http://www.projectsbyzac.com/1291/around-the-house-repairs-and-diy-fixes/upgrading-shop-lights-with-12v-dc-leds-using-352850505630-flexible-light-strips
Sorry, that should have been T8-style. I believe I bought them through Amazon.
DeleteJust a couple thoughts.
ReplyDeleteMany of the cheap LED replacements have really poor RF characteristics. I can't get anything on the radio below 5 mhz until I shut off the under cabinet lights and the down can replacements in the kitchen. My neighbor's motion activated driveway lights blow out all reception in my nearby bedroom. If you are a ham or SW listener, cheap switching psus are the enemy.
I've had the same experiences as others here- the curly bulbs never lasted as long as advertised, usually due to failures in their power supplies.
I've had a couple of LEDs fail here too. The one I tore apart and actually repaired (it's one of 13 matched fixtures with no matches available now) had a diode fail. That was easy to replace from my junk box and has been working fine since.
Like many things electronic, the failures in LEDs seem to be in the cheap power circuits, in caps, and exacerbated by poor heat management. All known issues, all put in the field anyway due to low price points...
nick
Oh, read labels carefully. Many bulbs and fixtures contain CREE leds, but the rest is build by the cheap china manf. No point in a CREE led if the caps fail...
CFLs and LEDs will suffer failure if they are used in a fixture with little ventilation, such as a recessed ceiling light. This is because they have electronic parts in the base that get too hot.
ReplyDelete