Glass globe failure of CREE LED bulb due to outgassing A21 100W, A19 60W

This is the glass globe from a 60W equivalent CREE brand LED light bulb. The silicone has been peeled off from the outside. The diffusing frost is from internal roughened glass surface.


I noticed that some of them become unfrosted. This behavior has also been noted by a few users here.

This is my CREE A21 100W equivalent 18W 1,600 lumen with a failed diffuser. It seems like an aesthetic only issue, but it is not. It results in diminished diffuser efficiency and increased glare in applications where the bulb is visible to the eyes. As you can see here, condensation of organic vapor from offgassing makes it clear enough to make out the LED emitters. It's a cause of concern for damage to LEDs and loss of output over the long run according to this report. http://www.lrc.rpi.edu/resources/newsroom/pdf/2012-2013/LED-VOCs8511.pdf



The fins along along the bottom gets up to about 200°F even in a totally open install. The heat causes something to outgas and condense onto the internal surface of the glass.




Just to prove my theory, I wetted the inside of the globe in the first picture with a bit of water. The effect of liquid on diffusing effect of frosted glass is devastating.

Does anyone remember a Buggs Bunny cartoon where he is the pitcher, throws an absurdly slow pitch that moves up/down in mid-air while the batter gets about 7 strikes on one pitch. That seems like this attempt at general purpose A-series bulbs by Cree.

I'm guessing that you got a bad bulb. Sounds like there is something wrong with the driver circuitry. Scrap it for parts, you just got sixteen nice XM-L leds for a fraction of their retail price. XM-L in bulk (1000 units) are $3 each, from a scrapped bulb they are $0.65 each.

I've been buying these bulbs since they were announced, probably two dozen of the 40W and 60W bulbs now. All have performed flawlessly. MichaelW, Cree scored a home run with these bulbs. They were the first bulbs under $30 to be dimmer-friendly and the first decent non-chinese bulbs under $10. Even now they are still cheaper than the chinese bulbs on dx. They've also forced Philips to produce a more affordable bulb as well. Though I am sad to see the 4Flow design, but maybe it'll help Cree stay in the black.

Lets see if we can count the strikes:
1: bad distribution from 'filament tower' design
2: unacceptable flicker
3: undersized heatsink (which they quietly attempted to address by slight redesign)
4: silicone coating failures
5: internal hazing
6: adhesive failure allowing glass dome to fall off
7: filtering to achieve higher CRI-that is just sad (but at least they did crank up the wattage, so output is the same. The 60 watt offering uses the 75 watt version's heat sink)
8: only 2700K & 5000K offering (but that is also a strike against most everyone else too)

The 3-way is nice. The low mode has sufficiently low consumption that you can leave it on overnight, the medium is good for timer usage, and high when you need light. Switch thought they would be the only one to fill that niche, not so.

Do you have any of these bulbs? Have you personally had any issues with them? These are the most popular consumer LED bulb, with great numbers comes the occasional quality control failure. Bright+ got a dud. Return it or dismantle it for parts.

Like I said, two dozen bulbs here. Mostly are heavily used and installed in places too unforgiving for CFLs. Zero problems with #1, #2, #3, #4, #5, #6 or #8. (My EDC is 3000K, I like warmer whites.) #7 I can't really comment on as I have not tried the "True White" bulbs.

MichaelW said:

Lets see if we can count the strikes:
1: bad distribution from 'filament tower' design
2: unacceptable flicker
3: undersized heatsink (which they quietly attempted to address by slight redesign)
4: silicone coating failures
5: internal hazing
6: adhesive failure allowing glass dome to fall off
7: filtering to achieve higher CRI-that is just sad (but at least they did crank up the wattage, so output is the same. The 60 watt offering uses the 75 watt version's heat sink)
8: only 2700K & 5000K offering (but that is also a strike against most everyone else too)

The 3-way is nice. The low mode has sufficiently low consumption that you can leave it on overnight, the medium is good for timer usage, and high when you need light. Switch thought they would be the only one to fill that niche, not so.

Click to expand...

1. No - it's not bad. I happen to like it
2. I don't notice any flicker except in photos.
3. Undersized? Based on what 'analysis'? 20 lamps in use at my place - no heat-related failures. (and you consider redesign to make a product improvement as a 'strike'? good grief)
4. Yawn - really?
5. Yawn - really?
6. Yeah - I agree with this one - I had one lamp with a loose bulb - not good.
7. You want high CRI in a cheap lamp - this is one way to do it.
8. Big deal - get over the fact that Cree doesn't make lamps in your favorite color.

parametrek said:

Do you have any of these bulbs? Have you personally had any issues with them? These are the most popular consumer LED bulb, with great numbers comes the occasional quality control failure. Bright+ got a dud. Return it or dismantle it for parts.

Click to expand...

#2: flicker bothers some people more than other. Photography is a big deal to many people.
I never had the glass fall off, but had one pop loose.

On the condensation of outgassing resulting in partial loss of diffuser effect, I had it happen to two CREE bulbs. Out of 5 or 6 I bought. I had it happen to a 9.5W year. and just had it happen to the 18 watter. I am not alone in this. Another member voiced his observation about "bald spot" appearing on his too. Since it's actually affecting its performance and noticeably increasing the glare, it's not a "yawn really?" cosmetic issues.

brickbat said:

1. No - it's not bad. I happen to like it
2. I don't notice any flicker except in photos.
3. Undersized? Based on what 'analysis'? 20 lamps in use at my place - no heat-related failures. (and you consider redesign to make a product improvement as a 'strike'? good grief)
4. Yawn - really?
5. Yawn - really?
6. Yeah - I agree with this one - I had one lamp with a loose bulb - not good.
7. You want high CRI in a cheap lamp - this is one way to do it.
8. Big deal - get over the fact that Cree doesn't make lamps in your favorite color.

Click to expand...

They made regular filament bulbs in linear/axial & ring shape for different distributions. This mimics the sidefiring axial, except with a rather large whole out of the top. Cree addressed this with the new 4Flow design (which substantially shrinks that black hole)
No, the fact that Cree changed the heat sink is an acknowledgement of the problem.
No one else is using filtering with LED bulbs, they using more premium LEDs
White is not a color, but an amalgam. To much blue in 5000K, and too much red in 2700K.

I understand that ownership told management, who then told accountants/MBA to develop parameters for a profitable consumer product, which then was sent as a rectangle for engineers to make a product that fits: cost/size/weight/durability.
This generation has ran its course.

MichaelW said:

No, the fact that Cree changed the heat sink is an acknowledgement of the problem.

Click to expand...

I'm pretty sure the heat sink was changed to reduce costs, not improve cooling or fix a problem. And the heatsink was changed very early, less than three months after the light was released. All of my revision one bulbs with the original heatsink are still going strong.

MichaelW said:

Lets see if we can count the strikes:
1: bad distribution from 'filament tower' design
2: unacceptable flicker
3: undersized heatsink (which they quietly attempted to address by slight redesign)
4: silicone coating failures
5: internal hazing
6: adhesive failure allowing glass dome to fall off
7: filtering to achieve higher CRI-that is just sad (but at least they did crank up the wattage, so output is the same. The 60 watt offering uses the 75 watt version's heat sink)
8: only 2700K & 5000K offering (but that is also a strike against most everyone else too)

The 3-way is nice. The low mode has sufficiently low consumption that you can leave it on overnight, the medium is good for timer usage, and high when you need light. Switch thought they would be the only one to fill that niche, not so.

Click to expand...

Name a better bulb,with as good specs,same or lower pricing, similar or better warranty.........you cant,period.

MichaelW said:

Lets see if we can count the strikes:
1: bad distribution from 'filament tower' design

Click to expand...

The couple of plots I've seen show it to be similar to a typical incandescent bulb.

2: unacceptable flicker

Click to expand...

This is indeed a problem with the Cree design if you notice 120hz flicker; not everyone does. This seems to be one of the bigger complaints about the Cree.

3: undersized heatsink (which they quietly attempted to address by slight redesign)

Click to expand...

This is only an issue if usage under nominal conditions produces unacceptable failure rates due to heat. From what I'm reading, these things are hanging in there with their contemporaries. It's a fine line - at the other extreme oversized heatsinks are bulky and weigh enough that some fixtures can't handle them.

4: silicone coating failures
5: internal hazing
6: adhesive failure allowing glass dome to fall off

Click to expand...

There are widely report failures, although rates are unknown to everyone but Cree.

7: filtering to achieve higher CRI-that is just sad (but at least they did crank up the wattage, so output is the same. The 60 watt offering uses the 75 watt version's heat sink)

Click to expand...

Depends on whether you need ~90 CRI or not. I just need "good enough" color rendering which all the LED bulbs I've bought since the Philips alien head have offered.

8: only 2700K & 5000K offering (but that is also a strike against most everyone else too)

Click to expand...

I suggest accepting that anything other than 2700K is presently just not going to catch fire with the residential market. 5000K is offered mostly for non-living space in the home and seems to sell at much lower rates than 2700 ... everything else is downright niche by comparison. A dozen people on CPF pining for something other than 2700 and 5000 does not a profitable demographic make.



It's a compromise. The failure of Switch suggests that there's little room in the market for a premium product that hits all the high points.

OP did you remove the silicone coating yourself? How did it come off?

To clarify to people, the fact you don't notice the flicker does not make it acceptable. Unexplained eye strain after hours of use in a reading application or computer area where it's the primary light and/or unexplained low grade headache would be the more likely result than any visible flicker.

Keep in mind they flicker as much if not more than a magnetic ballast fluorescent and you know how well loved those were.

Not disagreeing, Semiman, just want to learn more. Would you consider it too much of an imposition to point us to a recent study on this topic?

brickbat said:

Not disagreeing, Semiman, just want to learn more. Would you consider it too much of an imposition to point us to a recent study on this topic?

Click to expand...

Follow the links on this page:
http://www.led-professional.com/tec...to-avoid-stroboscopic-effects-in-ssl-lighting

slebans said:

Follow the links on this page:
http://www.led-professional.com/tec...to-avoid-stroboscopic-effects-in-ssl-lighting

Click to expand...

Per the summary presented in your direct link, these are studies of immediate perception of flicker. Only one link is useful to J.Random Public - Minimizing Flicker from SSL Systems; the other is behind a paywall.

Of more interest is the long-term effect of these systems along the lines of the fatigue, eye strain, and headaches that SemiMan is mentioning.

Thanks! - Skimmed through those. Interesting.

Where is the reference that links otherwise non-perceptible flicker to eye-strain and low grade headaches?

brickbat said:

Thanks! - Skimmed through those. Interesting.

Where is the reference that links otherwise non-perceptible flicker to eye-strain and low grade headaches?

Click to expand...

More info here:
http://apps1.eere.energy.gov/buildings/publications/pdfs/ssl/flicker_fact-sheet.pdf

A technical discussion here:
http://apps1.eere.energy.gov/buildings/publications/pdfs/ssl/poplawski_dimming_lightfair2012.pdf

You might consider searching the LED and Fixed Lighting forums as I have posted several links regarding flicker over the past 2 years.

OK - I've read all that. This is the closest I've seen to a basis for 'flicker causes headaches and eye strain':

"...and the flicker in magnetically-ballasted fluorescent lamps used for office lighting has been linked to headaches, fatigue, blurred vision, eyestrain, and reduced visual task performance for certain populations..." (from the USDOE Flicker Fact Sheet)

No citation to a scientific study - just a fuzzy-sounding statement from the US government - no author mentioned in that paper.

I'm not denying there might be a cause and effect relationship between non-perceptible flicker and headache/eye strain - I just want to read specifically about it.

brickbat said:

OK - I've read all that. This is the closest I've seen to a basis for 'flicker causes headaches and eye strain':

"...and the flicker in magnetically-ballasted fluorescent lamps used for office lighting has been linked to headaches, fatigue, blurred vision, eyestrain, and reduced visual task performance for certain populations..." (from the USDOE Flicker Fact Sheet)

No citation to a scientific study - just a fuzzy-sounding statement from the US government - no author mentioned in that paper.

I'm not denying there might be a cause and effect relationship between non-perceptible flicker and headache/eye strain - I just want to read specifically about it.

Click to expand...

OK - here is a small sample study by Philips. It contains references to the medical studies you are looking for.
http://lrt.sagepub.com/content/earl...45.full.pdf?ijkey=GcQ3UW7Qz2UwqtM&keytype=ref

I have spent a lot of time studying this issue mainly because I personally suffer from the stated symptoms. The sensitivity runs in my family - roughly 30% are affected while the remainder exhibit no symptoms whatsoever.