LED lights

[stoveliker]

This depends on whether the LED design actually works within standard net AC quality parameters and whether your AC falls within such parameters...

My point being that LEDs work fine elsewhere suggesting they can handle average power in most places, but not in your case, suggesting that your power is worse than normal. Worse than what it contractually should be?

 

[DBoon]

Just three weeks ago, I found our voltages running 138 and 265 VAC, for a full day before they re-adjusted.

I think this is likely the source of the problem. All the LED power supplies use a power conversion circuit (with MOSFETs) that will take your nominal 120 V circuit and convert it to a 500 Vdc bus that the power semiconductors switch at to control the light bulb, mostly for dimming. If the designer of that circuit is expecting (as they should) that the input voltage is 120V +/-10%, they are expecting a max voltage of 132 V.

6 V extra shouldn't make much of a difference, right? Not right. The ratings on the MOSFETs are nominally 600 V. It is possible that they design the circuit to have a 500 Vdc bus (for the MOSFETs) assuming a 120V input. 10% more means 550 Vdc bus, still ok. 15% more (138 V) means a 575 Vdc bus. Still technically under the limit, but getting closer to the 600 V nominal rating of the MOSFET, and you don't know what very short term surges or spikes might make things go higher than 138 V for a few milliseconds or seconds - your utility has no way to measure these very short-term spikes or surges, of course. They will just deny they exist. Power semiconductor MOSFETs have hard failure limits - they will fail completely after one very short duration event exceeding their voltage rating.

Basically, if your utility is delivering 138 V, you need to file a complaint (my opinion) as they are not operating within the +/-10% limit they should be operating within.

You might find success in an LED light bulb that doesn't permit any kind of dimming. It's possible that non-dimming bulbs might use a really cheap and dirty power conversion circuit that doesn't have MOSFETs in it, and therefore wouldn't be susceptible to the overvoltage problem.

 

[woodgeek]

Very "unclean", by comparison to other homes in which I've lived. We have a probably 600+ foot run at 220V to the closest transformer, and get voltage dips everytime an air conditioning compressor kicks on. Since we have four HVAC systems, that's basically several times per hour. On top of that, the autoranging function on the transformer on our street appears to not work properly, and I need to call the power company for a manual adjustment maybe once every 2-3 years. Just three weeks ago, I found our voltages running 138 and 265 VAC, for a full day before they re-adjusted.

But no matter, the incandescent and halogen bulbs take it just fine, and there's no good reason the LED's should not, as well. This is the reality into which they are being inserted, and engineers should take account of that reality, in their design.

Have you looked at a whole house surge suppressor? Or do you already have one?

 

[DBoon]

Have you looked at a whole house surge suppressor? Or do you already have one?

These are always a good idea to have (I have one at my house), but this will not be the solution to his LED light failure problem if the known voltage coming into his house is as high as 138 V.

 

[woodgeek]

These are always a good idea to have (I have one at my house), but this will not be the solution to his LED light failure problem if the known voltage coming into his house is as high as 138 V.

I agree that a surge suppressor will not do a thing to 138 V. But spikes are an alternative hypothesis here.

 

[ABMax24]

Or harmonic distortion could be the issue here.

 

[DBoon]

Or harmonic distortion could be the issue here.

The power conversion circuit converts everything to DC before turning it into pulse-width modulated (PWM) signals that drive the lighting circuit. If there is distortion on the AC input signal that does not impact the RMS voltage level of the AC input (nearly 100% likely) then distortion wouldn't have any impact at all. However, if there were surges of that were in excess of a few power cycles (100 ms or so), then I could see those having an impact on the lifetime of the power MOSFETs.

 

[gthomas785]

Most non-dimmable LED power supplies are not that sophisticated. The ones I've torn down (Edit: I should mention these were all the "filament" type) just put a bunch of diodes in series to bring the drop close to mains voltage and add a "current limiter" which is a resistor. There is typically no bucking or voltage regulation at all. If you buy the expensive bulbs, they'll have multiple such strings of diodes connected in opposite polarity so you get an apparent 120hz flicker instead of 60hz.

The current/voltage curve of a diode is not linear, obviously, so a small bump in supply voltage will result in greatly increased current through the diodes until one fails. I'm not surprised at all that the bulbs are quitting if the supply is 138V. That is in contrast to an incandescent filament which having (fairly) constant resistance would just run a few % hotter for the day.

You might have better luck with dimmable LEDs because they do at least contain some kind of voltage regulation, however I would not be surprised if those also failed under sustained over-voltage conditions.

 

[Ashful]

My point being that LEDs work fine elsewhere suggesting they can handle average power in most places, but not in your case, suggesting that your power is worse than normal. Worse than what it contractually should be?

They're working fine elsewhere? The internet is chock-full of customers complaining about the same unfulfilled expectations I just described. My experience is hardly unique. It's hard to gauge the numbers from internet chat forums, but it appears many people are seeing only a small fraction of the projected lifespan of these bulbs.

I think this is likely the source of the problem.

I'm not disagreeing that poor utility voltage control is a contributor, maybe even the leading contributor. I'm just saying that poor regulation from the utility is a reality, we cannot control it. My incandescent and halogen bulbs endure it just fine, they never fail in association with these events, and there's no justifiable reason a well-designed LED bulb assembly should not do the same.

Basically, if your utility is delivering 138 V, you need to file a complaint (my opinion) as they are not operating within the +/-10% limit they should be operating within.

Of course I do, each time it happens. They come out a day later, correct the issue, and it holds until the next major calamity. Then some other weather event causes them to switch configuration, and we're back to low or high voltage for a day or two, while I submit a new complaint.

I agree that a surge suppressor will not do a thing to 138 V. But spikes are an alternative hypothesis here.

We don't experience spikes, not that I've ever detected, nor have my smart UPS's ever recorded any appreciable spikes. We experience frequent (10x per day) voltage dips, but never spikes. The high voltage I mentioned has happened maybe 4x in 12 years, so average 1 day every 3rd year, or thereabouts.

 

[woodgeek]

I agree with all of the above, except the premise that internet reviews suggest many people are seeing short life for LEDs.

You are really behind the curve here @Ashful. The world is lit by LEDs and they work just fine. If you put one in a recessed fixture and it gets too hot, or a wet spot, then that is another story. But I am v skeptical that _everyone_ is getting 10-20% the rated lifespan of a mature technology used by billions of people. I want more evidence than angry and anonymous internet reviews.

 

[stoveliker]

The Internet is chock full of "bad stoves" too

I see no evidence in my life (including friends etc), was my point.

 

[Ashful]

But I am v skeptical that _everyone_ is getting 10-20% the rated lifespan of a mature technology used by billions of people. I want more evidence than angry and anonymous internet reviews.

I didn't mean to imply _everyone_ is getting the super-short lifespans I've observed in a few isolated cases, but the reviews make it clear some are. I would agree with you, those seeing such dramatically short lifespans must be a small minority. But I also suspect very few actually average on the advertised life spans (eg. 16 years @ 3-4 hours per day).

My location is definitely not wet. This pair is tucked way up inside a pair of large PAR38 cans in a porch ceiling, with at least 6 feet over overhang and a dropped soffit. No way have they ever seen moisture, other than some scattering blast when I pressure wash that ceiling once each spring. That was last done 11 months ago, in fact it's coming due, so unlikely a contributor.

I also can't imagine they're getting hot, as the cans are quite deep, vented, and designed to house halogen and incandescent bulbs up to something like 150 watt max. No way is a 7W LED putting off enough heat to nudge that can temperature more than a degree or three.

Maybe it was a fluke, and the other one will go another 12 years? The PAR14 LED's that I have around my adjacent porch seemed to exhibit that behavior, with one infant failure, followed by several years of no trouble from the remainder.

 

[stoveliker]

I think the statistics on any 16 year old LED would be very poor. Broad adoption of LEDs came later, if my memory is correct.

I.e. I think very few cases are to be made that LEDs reach that advertised lifetime. (But in my personal circle experience zero are having as short a lifetime as yours.)

 

[Ashful]

I think the statistics on any 16 year old LED would be very poor. Broad adoption of LEDs came later, if my memory is correct.

I.e. I think very few cases are to be made that LEDs reach that advertised lifetime. (But in my personal circle experience zero are having as short a lifetime as yours.)

Yeah, I'd believe that what I've reported here is extraordinary. I do have a few other LED's in locations that are doing better, such as my garage that's been gradually switched from 100W incandescant bulbs to A19 LED equivalents over the last 12 years. Of course, the garage bulbs only get a few minutes of usage per day, hardly the same as outside lights that run dusk until midnight everyday.

 

[begreen]

Back when I had incandescents installed I remember using 130v bulbs in the hard to reach locations. This is a trick I learned from commercial installations. They would go a very long time.

 

[Ashful]

The one that is often forgotten is traditional old T12 fluorescent tubes. I swear I must have some that are nearly 30 years old, still getting some daily use, and often all-day Saturday and Sunday in the shop. I've probably replaced fixtures more often than bulbs, swapping the old bulbs into new fixtures, on a few occasions.

I'd not want those ugly things lighting my house, but they're ideal for my work shop and electronics lab, as well as hiding over the door header inside most of our closets.

 

[stoveliker]

I just replaced all 16 of my fluorescent T8s in the basement with LEDs. I wanted to wait as I'm looking in redoing the ceiling (and the joist bays air sealing...), but they started to kick the bucket. Given the stove that was there likely when the basement was finished first, they could have been there since 1983. That's 40 years. Or they may have been 20 years or so.

LEDs tubes have better color though (I hate the bluish CFL light in my home).
Better color and half the wattage.

 

[begreen]

My friend changed out all of his T5 fluorescents in his woodworking shop last year. The difference is huge. Since changing he has more light, no radio interference, lower power consumption, and no cold startup flickering. I've been suggesting he do this for the past several years. "I should have done this years ago", he muttered the last time I saw him.

 

[gthomas785]

Yeah, In my unheated shop the fluorescent lights could sometimes take 15-20 minutes before putting out a useable amount of light in the winter. If I was running out there quickly to grab something there was no point turning the lights on, I would just bring a flashlight. The new shop LEDs are much brighter, higher CRI, more efficient, and most importantly for me, they start instantly in the cold.

 

[bigealta]

I changed the 3 single CFL 150W equivalent bulbs in my 2 car garage in utah to 250W EQ. LED (Sansi brand from ebay) single bulbs and it is a Huge improvement. I added 1 extra fixture to make a 4 bulb square pattern and the garage is now almost like daylight inside.I got a great deal on 4 bulbs for around $10 for all of them.

The upgrade could not have been easier.

They are 3000K which i prefer to 4000 or 5000 or even 6500.

I've since bought 8 more and have them in basement, closets, outside spots, etc. Anywhere i can get them.

 

[Ashful]

My friend changed out all of his T5 fluorescents in his woodworking shop last year. The difference is huge. Since changing he has more light, no radio interference, lower power consumption, and no cold startup flickering. I've been suggesting he do this for the past several years. "I should have done this years ago", he muttered the last time I saw him.

I never did the T5 thing, just T12's and T8's. The T12's definitely have better life, but the T8's I believe have better efficiency. I suppose both trends might continue toward T5?

Never had any issue with radio interference with any of my fluorescent fixtures, but I may be closer to the city, and it's admittedly been more than 15 years since I've even had a radio in my shop. Everything now is WiFi streaming to Bluetooth, which is never an issue.

I thought about LED shop lights when I redid the shop in 2015, and can't remember exactly why I went with fluorescent instead, but it might have had something to do with local available of fixtures in the lengths I wanted. In any case, my shop never gets below 55F, so cold starts aren't really an issue.

 

[gthomas785]

I thought about LED shop lights when I redid the shop in 2015, and can't remember exactly why I went with fluorescent instead

I know why. In 2015, LED lights were still a novelty. Kind of like electric cars now, where there's a real decision to be made with pros and cons to be weighed including cost, dependability, and limited choices.

In 2024, I don't believe anyone in their right mind would choose to install a brand new fluorescent tube fixture. I can understand the debate about incandescent vs LED with regard to color temp and CRI. But fluorescent? has no contest against LED. Sure, some people are stuck thinking LEDs don't work or are bad for this or that reason, but the technology has simply matured faster than the average human can adjust their mindset about things like this. Everyone will catch up eventually.

Keep your T12s going, there's no point junking them, but at some point (I predict within the next 10 years) those replacement tubes are gonna get hard to find.

 

[peakbagger]

I know why. In 2015, LED lights were still a novelty. Kind of like electric cars now, where there's a real decision to be made with pros and cons to be weighed including cost, dependability, and limited choices.

In 2024, I don't believe anyone in their right mind would choose to install a brand new fluorescent tube fixture. I can understand the debate about incandescent vs LED with regard to color temp and CRI. But fluorescent? has no contest against LED. Sure, some people are stuck thinking LEDs don't work or are bad for this or that reason, but the technology has simply matured faster than the average human can adjust their mindset about things like this. Everyone will catch up eventually.

Keep your T12s going, there's no point junking them, but at some point (I predict within the next 10 years) those replacement tubes are gonna get hard to find.

And harder to legally get rid of them. Our local town treats them like hazardous waste.

I remeber not that long ago that our mill had a 55 gallon drum with cover and home made crank handle with hole in the cover. Dead tubes would be inserted in the hole and the crank would be turned breaking the tube with the fragments dropping to the bottom. When it was full it went into the dumpster.

 

[stoveliker]

The mercury in those tubes is not good for anybody.
Some states (CA, ME, NH, MA, MN, VT, WA) prohibit CFLs in the dump. Others (NYS) do not prohibit this - for residential (small quantity) bulb waste. NYS does advise to bring them to a recycling point (blue or orange box stores take them for free).

So I brought mine to the orange store.

Large quantities of crushed bulbs (a drum would qualify as a large quantity, I surmise) generally has to be treated as hazardous waste. So That mill likely was violating the law, and putting mercury into the environment.

 

[peakbagger]

The mercury in those tubes is not good for anybody.
Some states (CA, ME, NH, MA, MN, VT, WA) prohibit CFLs in the dump. Others (NYS) do not prohibit this - for residential (small quantity) bulb waste. NYS does advise to bring them to a recycling point (blue or orange box stores take them for free).

So I brought mine to the orange store.

Large quantities of crushed bulbs (a drum would qualify as a large quantity, I surmise) generally has to be treated as hazardous waste. So That mill likely was violating the law, and putting mercury into the environment.

That was 30 years ago, before it was regulated.