Lights of America AC bulb replacements

[Dave_H]

Took a day off work, made the rounds of stores checking out
lighting...(among other things).

In Rona there were two Globe PAR30 spots ($19, which are made
up of a whack of 5mm LEDs) lit and pointed upwards but out
of reach. Neither was very bright, but one was noticably dimmer
than the other. They only offer one of this size, so they should
have started out more or less the same. No idea how long these
have run, but Rona has offered these for a year or less, and that's
less than 10k hours max.

On to Canadian Tire, saw the Sylvania power-LED spots (in the
package only). At $20 (for 2W,1 emitter) or $30 (4W, 3 emitter)
I'm not ready to jump in. Quality looks OK though with evidence
of heatsinking. I'll wait to see if they go on sale (missed one
recently).

Then to Costco, saw their LoA stuff. No wonder people get PO'ed
for having to buy two or three or four of anything instead of a
single unit to check it out first. I was able to buy the one and only
elsewhere at moderately higher unit cost, in a smaller size Costco
doesn't offer. They are the usual "showerhead" types.

I spoke to someone at Home Hardware who also sell LoA. They
are unaware of any recall. With Costco and Wal-Mart up here still
blissfully selling them , one has to wonder how long it will take for
the other shoe to drop.

Last week I hit Home Depot and noticed they are clearing the
Philips 1W coloured mini-bulbs for $1 each. Regular price is $10!
Not terribly bright but something to play with (blue, red, green)
(hey, RGB!). Yellow and multi-colour were not on sale, and I
didn't see white.

Dave

 

[Bullseye00]

The GE 2.0 & 2.5 watt LED bulbs ticked over 200 hours today and I did another comparison with the control bulbs.(Which only have a few minutes run time.) The GE bulbs that have 200 hours run time still seem slightly *brighter* than the controls, so they're definitely not dimming.(yet)
I got a couple of new lamps just to use for testing and started to calibrate with two new 10 watt CFL's, but I noticed a problem. The KillAWatt units are not reading the same voltage, one reads .4 volts higher than the other. they're plugged in to the same power strip coming directly from a wall outlet. I swapped them to opposite spots on the strip and the readings stayed the same.(.4 volts difference) It's less than 1%, but it makes me wonder if I should put off any serious testing until I can get better test gear.
After 1 hour I began observing and one meter ticked over to .01 kwh about a minute and a half sooner than the other. At two hours there was a little more than 3 minutes difference in the time they ticked over to .02 kwh. I'll swap the bulbs after 10 hours and see what the readings look like. It may be that the higher reading follows the bulb, meaning one bulb is out of spec, or maybe the spec just isn't that tight. Though it very well might be the meter, the one ticking over quicker is the one that reads .4 volts higher than the other.(Measured on the KilAwatt with nothing plugged, just sampling line voltage.) It may be that these cheap meters just aren't good enough for such sensitive testing. the alternative is to test using only one lamp and meter. But that will take twice as long, and at 100 hours per test will stretch into weeks.

 

[Dave_H]

The KillAWatt units are not reading the same voltage, one reads .4 volts higher than the other. they're plugged in to the same power strip coming directly from a wall outlet. I swapped them to opposite spots on the strip and the readings stayed the same.(.4 volts difference) It's less than 1%, but it makes me wonder if I should put off any serious testing until I can get better test gear.
After 1 hour I began observing and one meter ticked over to .01 kwh about a minute and a half sooner than the other. At two hours there was a little more than 3 minutes difference in the time they ticked over to .02 kwh. I'll swap the bulbs after 10 hours and see what the readings look like. It may be that the higher reading follows the bulb, meaning one bulb is out of spec, or maybe the spec just isn't that tight. Though it very well might be the meter, the one ticking over quicker is the one that reads .4 volts higher than the other.(Measured on the KilAwatt with nothing plugged, just sampling line voltage.) It may be that these cheap meters just aren't good enough for such sensitive testing. the alternative is to test using only one lamp and meter. But that will take twice as long, and at 100 hours per test will stretch into weeks.

Three minutes in 120 is 2.5% which is typically what you should expect for a consumer-grade device. AC measurements are inherently less accurate than dc, and instantaneous power is voltage times currrent which further increases error. Accumulating over time adds timebase accuracy error but this should be small if it is based on an internal crystal, or over long-term on the line frequency.

These devices are designed for larger loads, and our small loads are in the low end of resolution, therefore expect more error. My UPM-brand meter does not even state accuracy in its so-called "Spec" section which leads one to believe they don't want us to know how crappy it might be, before buying it. I haven't seen the KilAWatt spec, could you let us
know what is says (if anything)?

You might get more meaningful relative measurements using the same meter on all LEDs, which serializes the exercise (more time). Also, some type of calibration might work, but without knowing what the internal resolution is (independent of what is displayed), I doubt the value, other than to know how far off it is at some particular time and circumstances.

Nevertheless I read your experiments with interest (and might try
a few myself).

Dave

 

[Bullseye00]

At this point I kinda don't know what to do. Ideally, the 10 watt CFL's would tick over another .01 kwh each hour, on the hour. But the 10 watt CFL's are (once again) measuring slightly less than 10 watts, so they're ticking over that next .01 kwh later and later each time. AND the test rigs are out of calibration with each other, so one keeps ticking over later and later.(might be the bulb, might be the meter, or both)
I was thinking to try and compensate for the KillAWatt's lack of resolution that I'd observe the number of minutes each meter is taking to tick over each time, or make a final observation at the end.(Meter 1 took 612 minutes to get to 10 kwh, meter 2 took 615 minutes to get to 10 kwh.) But they're drifting further away from an even hour mark, and they're drifting further apart. So I'd have to stand there, being watchful and alert for quite some time. I thought about using a video camera, but my video camera can't see the LCD display of the meters, even fairly close up. So that won't work either.(dangit!)
But then I realized something that may make all this moot. In my previous testing, a 10 watt CFL came in at .98 kwh over 100 hours, and a 75 watt incan came in at about 7.8 kwh over 100 hours. Just a few percent off, one a little low, one a little high - that's not so bad. But the LEDs down in the 1 to 3.5 watt range are testing with much more variance from their rating. The 1.5 and 3.5 watt rated LoA lights came in at 67% and 63% of their rated power. Some cheap 1 watt rated LED bulbs came in at 50% of their rated power.(All tested w/ 100 hours run time.) So are all the LED bulbs consistently drawing *much* less power than their rating would indicate, or are the KillAwatt units increasingly *inaccurate* at lower power levels? Or is it something weird like the characteristics of the load of the LED bulbs is throwing off the accuracy of the reading? Or some combination of factors?
This leads me to wonder if I shouldn't stop my AC bulb LED testing until I can get better power metering equipment, and learn more about power metering.
With my time-lapse capable video camera I'm thinking of trying some "last man standing" type endurance tests for may various LED flashlights and lanterns. I can set them all down in frame with a clock, turn 'em all on, start the camera and just let it all go for a few days. The only problem I can foresee with that is that over time all the lights dimming might throw off the camera's autoexposure and mess up the visual accuracy of the video recording. I might be able to resolve that if I keep the room lighting consistent, or fiddle with manual exposure. I'll have to do some preliminary tests and see. So I could keep busy with other testing.
But I do need to figure out what to do about my AC bulb testing. I might buy a more expensive meter, but I'd probably only get one. That would make testing a lot longer and also make me wonder about the overall accuracy. Here in Central Florida, the land of bad power and wild temperature swings, I'd wonder about the consistency of each of the tests conducted at different times. I can look at the KillAwatts and watch the voltage swing up and down by a few points over a few minutes. And they're both reading about half a volt higher today than they were yesterday. The temperature can also vary widely during the day, and from day to day especially in (what we call) Winter. You can go to work in a heavy coat and need to change to short sleeves by early afternoon. Or you can have several days when you need a heavy coat, then a day or two when you're wearing short sleeves. I'm monitoring indoors, but the temperature is still going to vary, especially when I'm not home to use the heat and/or air conditioning to maintain a (fairly) constant temperature. And, to top it all off, the KillAwatt units that I have don't have memory.(The newer model does.) So if there's a power outage, I'm having to start that particular test over.
I don't know how much overall the minor variations in line power and temperature will affect readings. It's not like I'm monitoring a nuclear reactor or doing cancer research, but I would like to be as accurate as possible. It'd be much better(I think) to be running all the tests at once, on the same AC circuit. But that'd mean buying 4 or 5 expensive meters.

<sigh> Nothing is ever simple.

 

[jtr1962]

I've tested a large number of LED bulbs and might have some insight into the problems you're facing. First off, most of them are unregulated, so a fluctuation in voltage of a few volts in either direction can affect power consumption by 10% or more. This probably accounts for the inconsistent readings. Even if a lamp is regulated, often the convertor's efficiency changes with input voltage, so again the input power will vary with the voltage, although not as much as in the unregulated case.

Second, yes, the resolution of the Kill-A-Watt is poor, so what I usually do is read the VA (which is usually higher than the watts unless the power factor is 1), and multiply by the power factor to compute the watts. Since most LED bulbs have power factors under 0.5, this gives you about twice as much or more additional resolution.

Third, I'm not really sure how the Kill-A-Watt computes kW-hr. If it simply takes the reading in watts, and mutiplies by time, then you end with a certain granularity to the kw-hr readings. For example, a load which barely reads 3 watts and load which is almost 4 watts, but not quite high enough to read 4 watts, might both produce the same kW-hr reading. Or perhaps the Kill-A-Watt uses tenths of a watt when computes kW-hrs although I doubt it. That would imply at least a 16-bit A/D converter which I doubt the unit has ( 12-bit is more likely). So in effect you get no more resolution letting the meter tally kW-hrs than you do just taking the instantaneous reading.

Fourth, perhaps we should contact the manufacturers of the Kill-A-Watt with our concerns. It might be possible to get the unit to read to tenths or even hundreds of a watt without too much trouble. You just need some sort of autoranging. Instead of having one scale which goes from 0 to ~2400 watts have several ( 0 - 2400, 0.0 - 240.0, 0.00 - 24.00 ). No need for a higher resolution A/D convertor, and probably just a need for a few extra sense resistors plus a firmware change. Come to think of it, maybe it might be possible to mod a unit by changing out the sense resistor with one 10 or 100 times higher. Naturally, the unit would now be limited to lower wattages, and the displayed reading would be either 10 or 100 times the actual power consumption, but in theory it should work (unless the unit uses something other than a sense resistor to measure current).

 

[usLEDsupply]

I've tested a large number of LED bulbs and might have some insight into the problems you're facing. First off, most of them are unregulated, so a fluctuation in voltage of a few volts in either direction can affect power consumption by 10% or more. This probably accounts for the inconsistent readings. Even if a lamp is regulated, often the convertor's efficiency changes with input voltage, so again the input power will vary with the voltage, although not as much as in the unregulated case.

Second, yes, the resolution of the Kill-A-Watt is poor, so what I usually do is read the VA (which is usually higher than the watts unless the power factor is 1), and multiply by the power factor to compute the watts. Since most LED bulbs have power factors under 0.5, this gives you about twice as much or more additional resolution.

Third, I'm not really sure how the Kill-A-Watt computes kW-hr. If it simply takes the reading in watts, and mutiplies by time, then you end with a certain granularity to the kw-hr readings. For example, a load which barely reads 3 watts and load which is almost 4 watts, but not quite high enough to read 4 watts, might both produce the same kW-hr reading. Or perhaps the Kill-A-Watt uses tenths of a watt when computes kW-hrs although I doubt it. That would imply at least a 16-bit A/D converter which I doubt the unit has ( 12-bit is more likely). So in effect you get no more resolution letting the meter tally kW-hrs than you do just taking the instantaneous reading.

Fourth, perhaps we should contact the manufacturers of the Kill-A-Watt with our concerns. It might be possible to get the unit to read to tenths or even hundreds of a watt without too much trouble. You just need some sort of autoranging. Instead of having one scale which goes from 0 to ~2400 watts have several ( 0 - 2400, 0.0 - 240.0, 0.00 - 24.00 ). No need for a higher resolution A/D convertor, and probably just a need for a few extra sense resistors plus a firmware change. Come to think of it, maybe it might be possible to mod a unit by changing out the sense resistor with one 10 or 100 times higher. Naturally, the unit would now be limited to lower wattages, and the displayed reading would be either 10 or 100 times the actual power consumption, but in theory it should work (unless the unit uses something other than a sense resistor to measure current).

i haven't opened up a Kill-A-Watt yet but in most cases the ac current is measured with a coil around the hot leg so if you open it up and find a coil around the wire you should be able to take the wire that runs threw the coil and loop it threw again then you should have twice the sensitivity if it reads 10watts you would only be using 5 watts? (for every time the wire goes threw the ring your sensitivity should go up)
it's just a guess but it could be a simple fix (that is how we do it on whole house systems)

as for the replacement LED lights, here are some pointers

1: don't but any led bulb made up of the small 5-8mm led's and expect it to be bright (more then 25w equivalent)

2: brightness and wattage go hand in hand

for MR-16 anything less then a 6w LED will put out less light then a 50w halogen so don't expect to be able to light your kitchen with a 3watt led light

for EU-27 (standard screw in bulb) a 5watt LED is going to be in the 40-60watt range and a 10-14watt LED is going to be in the 70-100watt range

you will find the 3x1w Cree led lights to be nice for areas you don't need a lot of light but they will not put out like a 50w bulb i don't care what the packaging says.

as for the color temp they have led's that put out a nice warm white (almost too warm) and my favorite the netural white and the standard cool/blue white but the cooler the white the more light output you are going to get for the same wattage so pick your trade-offs
for example our 5w Cool White Bulb outs out 350lm and our 5w Warm White Bulb only 300lm

 

[Bullseye00]

Thanks to everyone who has responded with support and advice. I was hoping that I'd hear from people with more electronics knowledge and experience than I have.(which doesn't take much)

re: Line voltage variance - I know the line voltage is high here. A friend of mine measured it with a decent meter years ago, so I know that the KillAWatt is close when it reads within a few tenths of 125 volts. I can watch the voltage measurement slowly swing up and down 1 or 2 tenths at a time, varying gradually over the course of the day. They were measuring in mid and high 124 volts last night, high 124 and low 125 earlier today, and now they're back down in mid and high 124 again.(fairly consistently .4 or .5 volts apart) The odd thing is that they don't generally vary at the same time, or even in the same direction. They just kinda bump up and down independently of each other, though they're plugged into the same outlet, via the same power strip. I can understand one reading higher than the other.(bad calibration) But how can they *not* vary at the same time, or in the same direction? All of the other measurements, HZ, Amps, VA, watts(with the 10 watt CFLs, KillAWatt can give me a watt measurement) are steady, except for what P3 calls "Power Factor".(I'll have to look that up in P3's docs and see what exactly it is) Power factor jumps slowly up and down, probably varying with voltage. They're both measuring the Hz as 59.9. I don't know if they're both off by .1 Hz, or if the line voltage is actually off by .1 Hz.(How likely is that?)

As for modding a KillAWatt, I've soldered before but I don't have the dexterity or skill to do work on a circuit board as small and tight as I'm sure that one is. Regarding suggestions to P3(the company that manufactures the KillAWatt) I can see where they might be receptive to increasing the sensitivity and accuracy in the lower range because more low power devices are being produced and they'll want consumers to be able to measure how much their new devices/appliances are saving them. It'll all depend on getting the to the right person at P3.(Which will involve luck more than anything, I think.) But I don't think that they're going to want to add much complexity to the product. It's very much a consumer rather than semi-pro product line. The newer model(than mine) is more of an *upgrade* than an *update*. It looks like the sensitivity is the same, and they've only added a memory and the ability to input your local rate/kwh so that the device can automatically tally the power used and keep a running total of how much that device costs you to run. I've gotten some good advice and I'll try to work around the limitations of what I have.
The first thing I thought I'd try is to swap the CFLs and see if the lower power measurement stays with the meter or follows the CFL. Then I'm going to measure a very low wattage incan bulb to see how the KillAWatt measures a simple load down in that range. I've found 4 and 7 watt incan bulbs, and I know I can get 4 and 7 watt LED bulbs locally. I don't think I'll be able to get incan bulbs lower than 4 watts, and I don't think I've seen any CFL's down in that range(under 10 watts) either. But I'll start with the low range incans and see what happens, then proceed from there.

As for bulbs with many small(3-5mm) LEDs, I'm not going to trust those any more. I've had first hand experience and know how bad they can be. And yes, I'm also somewhat suspicious of LED bulbs in the 60-100watt(equivalent) range. I can believe they'll last longer than the bulbs with arrays of small LEDs, but I don't think they'll put out as much useful light as the manufacturers claim they do. I'd like to try something like a Zetalux, but I'm going to see how these GE 2.0 & 2.5 watt (single emitter) bulbs do first. They seem to be holding up so far, and they're way cheaper than LED bulbs with a higher power rating. And if they last me a year or two, there'll surely be something better out then. Being an early adopter, I'm not concerned that these early products last for decades. The tech is relatively new, and we're going to see a lot more advancements in the next few years. Now is actually a pretty bad time to settle on lighting tech that you'll use for decades to come. Hopefully the commercial and home lighting industries will come up with new standards for LED lighting and we'll see homes and buildings fitted with new lighting fixtures not locked in to the decades old "light bulb" standard sockets.

 

[Bullseye00]

I swapped CFLs and the meter that read high before still read high. So the problem is with the meters or the lamps. And considering that meter which reads the line voltage as .4 volts higher with nothing plugged in to it is the one that's giving higher power readings, I'm inclined to think that it's a meter problem.
Regarding the possible inaccuracy of the KillAWatt units at low power(10 watts and under) vs possible issue with load characteristics of CFLs and LEDs, I've put a 4W incan in one lamp, and a 10 watt incan in the other. At the 50 hour mark the 10 watt incan was reading .46 kwh, which would work out to .92 kwh over 100 hours. Pretty much in line with what the 10 watt CFLs were reading at 100 hours. The 4 watt incan looks like it will end up being about .3 kwh over 100 hours. Which would be a significant deviation form its rated power. So that's looking like it may be an issue of increasing inaccuracy the further you go under 10 watts.
The good news is the GE 2.0 & 2.5 watt LED bulbs still seem to be holding up well after 300+ hours. The bulbs that have run over 300 hours still seem to be slightly brighter than the bulbs with only a few minutes run time. I am cautiously optimistic.

 

[ryball]

I don't know if it has been mentioned, but I got a letter from Costco that said their records showed that I had bought one and they had a ton of complaints about the LOA LED lights that were sold there. It went on to say, if I was unsatisfied for any reason, to bring it back to get a full refund.

FYI.

 

[Dave_H]

I don't know if it has been mentioned, but I got a letter from Costco that said their records showed that I had bought one and they had a ton of complaints about the LOA LED lights that were sold there. It went on to say, if I was unsatisfied for any reason, to bring it back to get a full refund.

FYI.

Costco Canada is still selling the 3.5W round globes at 2/$24
and 1.5W chandelier style at 3/$19, so any recall hasn't made
its way through. My membership is lapsed so couldn't buy any,
anyway. Costco was always good about returns, if you were
reasonable, but the number of returns might be small enough
that they don't know what's looming.

I decided for the moment to continue to run my only LoA
bulb outside as a porch light, to see if cold ambient has any
effect.I don't remember if anyone commented on whether
cooling the bulbs will likely improve the meager half-life,
and if the failure mechanism is thermally-related.

Dave

 

[Bullseye00]

re: if the failure mechanism is thermally-related - I haven't really explored this much. I did run one without the clear bulb surrounding the LEDs. I figured this might allow it to run cooler because the air wouldn't be trapped around the LEDs and there might be some minor convection. But that one had the LEDs dim also, taking a similar amount of time. Note that this was during Summer in Florida, so the temperature was fairly warm.(In the 80s most of the time.) I haven't tried to fan one or employ any other method of active cooling. I was thinking of trying to heat sink the LEDs, but that's way too much of a project for such a cheap product. Let us know how the one on your porch does.

re: recall/refund - Sam's Club doesn't appear to know about the inherent issues with this product. I'm thinking of going in and telling the manager that he should stop carrying the LoAs, or people may quit buying the GE LED bulbs as well. And I'm sure the markup on LED bulbs is enough that he'd want to keep selling them.(the ones that work anyway) But I'm going to endurance test GE bulbs for a while longer before I make that recommendation.

 

[Dave_H]

On the subject of GE, I've just seen 1W GU10 spots in a department
store (Zellers). They're similar to LoA and Globe products, made up from
a bunch of 5mm LEDs. At about $18 they're twice the price of the others.
Looks like the store doesn't have the proper price point figured out. I
could get a single-emitter Sylvania 2W spot for a few bucks more elsewhere,
maybe less if/when they go on sale again.

Is there any reason to believe the GE spots are likely to be any better
than the others i.e. besides the GE name?

I didn't get a product number, will do next week.

Dave

 

[Dave_H]

I haven't tried to fan one or employ any other method of active cooling. I was thinking of trying to heat sink the LEDs, but that's way too much of a project for such a cheap product.

Agreed, there is only so much one can do at this level. I might
disassemble it (if not returned), any suggestions about detaching
the glass bulb cleanly?

Another observation, these have a distinct afterglow after being
switched off. I presume the ballast contains a capacitor, and it
is just charge leaking away; or is it the phosphor (never seen this
with any other LED products)?

Your summer ambient is about 50F (28C) higher than current
temp around here (which will get colder). I know in general for
semiconductors the failure rate decreases roughly 10x for every
10C lower temperature, but I would say it does not necessarily
apply to this type of failure, or even LEDs in general. But if
thermally-related, some life improvement is possible, though
unlikely to get it to the cost breakeven point (using 300 hours
as a baseline)

Dave

 

[Bullseye00]

re: detaching the bulb cleanly - Yes, this can be done. And without tools - but BE CAREFUL. Before I realized that these wouldn't last, I tried to get the globes frosted. A friend of mine is a machinist and thought we might try sand blasting the inside. He just took each of them in his hands and (somehow) deftly removed the globes. But, we learned that the globes are fairly thin and fragile. Sand blasting them didn't make much of a frost, and it did make the globes *very* fragile.(One crumbled in his hands while he was sand blasting it, leaving lots of sharp points and edges.)
I don't know how he did it, but I tried removing globes from a couple of my LoA bulbs. I tried twisting a globe off and it separated from the ring that attaches it to the base. Then I took another one and rotated my wrists(prying up on one side) and the globe popped off.(With the ring still attached, so it ought to be able to go back together.) If you try to remove the globe from one of yours, *please* use heavy gloves or something to protect your hands.

re: temperature related failure & afterglow - Hopefully someone with a strong electronics background will jump in and enlighten us about this.

re: GE 1W GU10 w/ bunch of 5mm LEDs - I wouldn't trust anything with an array of 3 to 5mm LEDs. I'd look for something with 1 to 3 large emitters. The GE 2.0 & 2.5 watt bulbs I'm currently testing are holding up well. You might try finding some of those. They aren't as directional as the GE 7 watt spot which seems to be well received here on these forums.(Also my parents have been using a pair of these for months, and they still seem to be holding up well so far.)

re: KillAWatt - I have the 100 hour numbers on the 4 watt & 10 watt incan bulbs. The 10 watt incan came in at .92 kwh, very close to the 10 watt CFLs at 100 hours, at 92% of its rating. The 4 watt incan came in at .31 kwh, which is 77.5% its rating. With incan bulbs, I think we can rule out odd load characteristics. And the 10 watt incan measurements are within a couple of percent of the CFL measurements. So it looks like the KillAWatt is increasingly inaccurate the lower you go below 10 watts. The 7 watt bulb will probably come in somewhere in the 80's. Just for the sake of curiousity I've started that one on a 100 hour run and should have the results in 4 days. But it's looking like the KillAWatt is not accurate enough for testing devices under 10 watts.

 

[purduephotog]

I bought two myself for using in the bathroom. I thought that they would pay for themselves quicker that way as bathroom lights are often turned on and off in about a minute which isn't good for fluorescents. I thought with short term use, even heatsinking wouldn't be much of an issue. I was wrong about those bulbs. With temporary use, one has burned out already with hardly any use at all. These things are pieces of junk. If you want to use l.e.d.s for home lighting, get some good ones with Luxeon Rebels or Cree XR-Es. Even better would be to wait until they have Cree XP-Gs in a bulb. Right now, the vast majority of l.e.d. lights for this application are either way to expensive, are cheap pieces of junk that don't last, or at best have the same efficiency as cheaper CFLs.

Same here. I have two- one is down to 3 working LEDs in less than a year. The other (still brand new) suffered a single LED failure the one time I plugged it in. It is about 8x brighter than the old one.

Don't waste your money.

 

[Dave_H]

I just bought a Sylvania 2W A19 bulb from Canadian Tire for $13;
actually closer to $15 after taxes. Ironically, CFL's are exempt from
8% Ontario sales tax, but not LEDs...

The part number is nowhere marked on the package except above the
UPC barcode; it is LED2A19/F/830/HVP (Accent Series LED).

Through the package in the store I could see it was not a "showerhead"
multi-5mm bulb, apparently a single 2W emitter, but no external
heatsinking. The lightly-diffusing bulb appears to be molded
plastic; disassembly not planned at present but looks like a hacksaw
might suffice.

Colour is "soft-white" 3000K, rated 50 lumens, life 15,000 hours to
70% brightness (not great). Light is OK so far but has artifacts on the
ceiling above (bright band etc.). UPM power meter reads 2W. Testing
will continue.

Dave

 

[LEDninja]

This one?

 

[Dave_H]

That is it. I thought to post the link, but a bit unclear on
the site policy for this; otherhand, others are posting similar
links.

I've run this bulb for several hours, and the collar between
the bulb and socket gets comfortably warm, not hot. That
does not give accurate temperature of the baseplate or
the die. I have a thermocouple-type temp. sensor which
can be attached to the outside to give some measure of
temperature rise.

Minor observation, there seems to be about a half-second
delay from switch on to light on.

Dave

 

[LEDninja]

In general advertising is not allowed. Mentioning an item together with a link is considered advertising.
Links provided for clarification purposes is tolerated. It is best to 'hide' the link - see how I changed mine.

That is it. I thought to post the link, but a bit unclear on
the site policy for this; otherhand, others are posting similar
links.

That is OK.
I have some bulbs that goes very bright, then dim down in half a second. That is worse as the LED is overdriven at every startup.

Minor observation, there seems to be about a half-second
delay from switch on to light on.

 

[Dave_H]

re: detaching the bulb cleanly - Yes, this can be done. And without tools - but BE CAREFUL.

snip

One crumbled in his hands while he was sand blasting it, leaving lots of sharp points and edges.)

snip

If you try to remove the globe from one of yours, *please* use heavy gloves or something to protect your hands.

Your concern appreciated, I'm not out to slice into my hands
or any other body parts. I was just thinking that for glass globes
that don't want to detach, would it work to score the glass around
the neck with a glass cutter (or something hard/sharp) then
heat, then cool the globe by immersing it partly in ice water
(not the socket or electronics) such that thermal stress would
cause a clean fracture? I might try this with a cheap incan.
first. It might however cause the whole globe to fracture if
it is so thin as the LoA in question is.

Dave