Cree A19 9.5w 60w 800lm 2700K for $13.97

[lildave]

I work for a major company and we have connections.

 

[lildave]

We sometimes get prototypes.

 

[idleprocess]

They look and work great. But they unsolder themselves if they get hot ( ie Jelly jar). I just replaced 30 at work.

That's a bit disappointing - especially with hotter-melting RoHS solders being all the rage in industry these days.

 

[lildave]

If they're not in the jelly jar they'll last (I hope). In the jar they can't dissipate the heat.

 

[bose301s]

I work for a major company and we have connections.

We sometimes get prototypes.

Well, I don't think prototypes should be used to judge any final product to be honest.

Also, what company do you work for, I just don't see these getting out to any company other than Home Depot.

 

[BVH]

OK, who was the 1st CPF'r today to ask about availability at the San Luis Obispo Home Depot? I was the second. Employee said I was the second one in a couple of hours to ask. Wouldn't think John Q Public would be asking so figure it's gotta be another CF'r

 

[JohnR66]

I'm happy to see this. After the 40w (equiv.) Utilitech bulbs of two years ago on sale for around $10 and some interesting products of 1 year ago, I figured 2012 was the year for LED bulbs, but it never happened. Still lots of high cost products and much of it around CFL efficiency (~60 lumens/watt).

I predict that later in the year, we will see a lot of new higher efficiency bulbs at under $15 now that Cree has raised the bar.

 

[dirtytoenails]

Actually, my calculations were based on using XT-Es, not XP-Es. I fixed the typo in my post. XP-Es are inherently much more costly than XT-Es. You can definitely get the worst bin XT-Es for slightly under $1 in bulk (5000s). We're obviously dealing with much larger quantities here, and Cree avoids two sets of middlemen, themselves and the distributor, which would increase the costs for anyone else making a lamp like this.

Thinking about the price, a possibility is both Cree and Home Depot might be selling the lamp at cost just to undercut everyone else. Or perhaps Cree is using the sales of bare LEDs to cover for short-term losses in the A19 retrofit market. Either one might actually be a smart business strategy. By the time everyone else loses a significant market share, Cree's production costs will have dropped so that it might make a good profit even lowering the retail price to $10 because it has volume on its side. There will be plenty of niche markets for A19 lamps with high CRI, or CCTs other than 2700K or 5000K, for the other manufacturers. Cree seems to be interested mainly in the high-volume products.

So I received mine today, measured them, tore them open, and found what made them tick. I think Cree and HD are selling these close to cost based on the insides (so sell your stock as they can't be making enough profit to justify their P/E ratio!).

The 60W warm white has 20 LED packages. Each has 4 die mounted to an aluminum nitride substrate of the XT-E footprint but with different circuitry. This seems to be something they only make for themselves (and it'd probably make a poor solution for anything other than stuffing behind a diffuser). The die are about 610 um x 610 um, making each package have 1.44mm^2 of die and the total lamp die area being about 29mm^2. That's a lot of epi and is probably about 3 bucks. Add on the package substrate, the phosphors, the silicone lens, the manufacturing and test, etc. and these things have to be north of $10 to make. Then you package them, pay UPS freight to get them to HD, stock them, sell them to me at $12, ship them to me in CA from MD, and you can't have made any money on this... anyone know if they are using this in their new troffers?

The 40W and the 60W each have 10 XT-E HVW 24V packages in them. They have 8 DA3547 die per package, or ~1.3mm^2. Total lamp epi is 13mm^2, or slightly more than an MT-G2. Again, they've utilized the XT-E footprint but with different circuitry. All LEDs are in series, meaning one die opens up and the whole thing goes dead.

None of the lamps had stellar efficiency LEDs in them given the low drive current, but the driver was freakin' incredible. It outputs 224VDC to the LED strings at 93% efficiency! If you bust open your globe, be careful as you're working with some dangerous voltages here. Lots of electrolytic caps and the driver didn't look all that inexpensive to make. Probably $2? Reliability should be interesting for these, especially in enclosed fixtures.

No mention of useability in enclosed fixtures on the packaging.

Anyone else have any fun with these?

 

[idleprocess]

After the 40w (equiv.) Utilitech bulbs of two years ago on sale for around $10 and some interesting products of 1 year ago, I figured 2012 was the year for LED bulbs, but it never happened. Still lots of high cost products and much of it around CFL efficiency (~60 lumens/watt).

I got one of those Utilitech bulbs nearly 2 years ago for $10. It has been going steady 24x7 ever since in a semi-enclosed fixture with perhaps 48 hours total downtime (power outages, occasionally switched off momentarily, once or twice for ~8 hours).

That light is my yardstick. I expect LED bulbs to all perform about that well or better.

 

[LEDAdd1ct]

The 60W warm white has 20 LED packages. Each has 4 die mounted to an aluminum nitride substrate of the XT-E footprint but with different circuitry.

The 40W and the 60W each have 10 XT-E HVW 24V packages in them. They have 8 DA3547 die per package, or ~1.3mm^2.

Is it the case that the 60W warm white has 20 LED packages, and the subsequent portion stating that the 60W has 10 packages is describing a 60W cool white version?

 

[dirtytoenails]

Is it the case that the 60W warm white has 20 LED packages, and the subsequent portion stating that the 60W has 10 packages is describing a 60W cool white version?

Yes, that's right.

 

[jtr1962]

None of the lamps had stellar efficiency LEDs in them given the low drive current, but the driver was freakin' incredible. It outputs 224VDC to the LED strings at 93% efficiency! If you bust open your globe, be careful as you're working with some dangerous voltages here. Lots of electrolytic caps and the driver didn't look all that inexpensive to make. Probably $2? Reliability should be interesting for these, especially in enclosed fixtures.

This design actually makes sense from an electrical standpoint. First run everything in series so you don't have to worry about balancing current among strings. Second, instead of using 20 LEDs with a total Vf of ~60V, use more dice in series to get the total voltage higher and current lower. Both things lend themselves to greater driver efficiency. It's more efficient to slightly step up (or down) the ~165 VDC available from rectified, flitered 120 VAC than to drop it to 60 VDC or less. My guess is if dimming wasn't required efficiency could have easily been in the high 90s. As it is 93% efficiency is the highest I've seen for a dimmable offline driver.

The only downside is if one LED fails, the entire lamp goes dark. Cree probably ran the numbers for one die out of eighty failing, then ended up with a probability close to zero during the 25,000 hour rated lifespan.

 

[wws944]

OK, who was the 1st CPF'r today to ask about availability at the San Luis Obispo Home Depot? I was the second. Employee said I was the second one in a couple of hours to ask. Wouldn't think John Q Public would be asking so figure it's gotta be another CF'r

The CEO of Cree was on Fox News. I linked to it above. So many people would have found out through them - even here on the Left Coast.

I was the first to ask at my Home Depot. (None in stock.) The gal stated that HD was coming out with a lot of new LED products this spring.

 

[Harold_B]

Any idea of what happens with a Cree die when it fails catastrophically? Does it fail to open or dead short? If it's the later then the bulb will still work.

 

[jtr1962]

Any idea of what happens with a Cree die when it fails catastrophically? Does it fail to open or dead short? If it's the later then the bulb will still work.

It's morely like that it'll fail as a dead short due to the high power supply voltage. First the LED open circuits. Because no current is flowing in the other LEDs you essentially have the entire power supply voltage (~224VDC) across the open LED. This will arc across it and probably make a conductive path across the die, and then the entire string will light again. It seems to work this way with strings of LED christmas lights, where more often than not individual LEDs short when they fail.

 

[Marcturus]

CEO interview's automatic transcript at Fox -
"to give you a three minute commercial and Avago lecture"

 

[carbonita]

This bulb has the right form factor of A19 for a dining room chandelier (Poul Henningson PH4) I have and is dimmable. The CRI is reasonable at 80+, and is dimmable. But I've gotten other dimmable LED bulbs that don't dim enough before popping to zero lumens (ecosmart PAR30's) even when paired with suggested dimming switches.

Anyone find the dimming level for these bulbs?

 

[bose301s]

This bulb has the right form factor of A19 for a dining room chandelier (Poul Henningson PH4) I have and is dimmable. The CRI is reasonable at 80+, and is dimmable. But I've gotten other dimmable LED bulbs that don't dim enough before popping to zero lumens (ecosmart PAR30's) even when paired with suggested dimming switches.

Anyone find the dimming level for these bulbs?

Down to 5%.

 

[EscapeVelocity]

Nice!

 

[SemiMan]

They look and work great. But they unsolder themselves if they get hot ( ie Jelly jar). I just replaced 30 at work.

That would take over 200c .... I would expect other failure modes first. Are you 100 % certain they were made by Cree an not just had Cree LEDS in them? Can you post pics?

Semiman