New LED manufacturer in Silicon Valley

Scottiver

Enlightened
Joined
Jul 7, 2005
Messages
585
Location
California
Just read an article in the SF Chronicle business section about a new LED manufacturing start-up company that's going to be making some new highly efficient LED's. Their focus seems to be replacing light bulbs at homes and businesses with these new emitters but perhaps their new technology will trickle down to the flashlight world. Interesting article though.
http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2012/02/08/BU7M1N48S9.DTL
 

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
Interesting. Don't know that I'd call what they are doing as "highly efficient" or not but it is interesting. They say it is only 4x more efficient than incandescents. Not that much these days. They say they can run hotter and "produce more light per area than any other LED on the market". That last statement can be taken multiple ways so I will reserve my excitement until I can check out the LED myself. My guess is they are woefully short in the efficiency game but make that up by driving the LED super hard. Notice a conspicuous lack of lumen data? Then they try to boost the marketing claims of a short 1 year payback period based on 24/7 operation. I am going to be getting one as soon as I can to examine it in detail and see how it truly compares to current LED tech.
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
MR16 format, using "1/4 the power of incandescent lights."

So...13W MR16 LED? Is this news? Anyway, every combination I tried n the "Find your Light" product selector on their web page said "Specs coming soon." Let's hope they come to more than this! Is it such a good idea to go public with press releases without any meat on the web site? I mean, I could infer output from 2400 cd and the spot widths they quote, but that's not really so accurate.

Actually, I found a spec sheet:

PDF clicky
They say that they're rated in recessed but not enclosed fixtures, and want ambient temperature below 45C. The lights also apparently shut down (or dim? It's unclear) when overheated. I think this means that your fixture enclosure had best not get over 45C/114F.
 
Last edited:

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
They say they will start by selling MR16 bulbs to 'showcase their ability to take on challenging lighting market segments.'

Other than the marquee' over the counter at KFC that's not a very big segment. :)

And no, I'm not being sarcastic (I wish I were). That's the last place I saw LED retrofits used in a track scenario in a commercial environment. While I've always liked the MR16 format for halogen, especially those marvelous Solux, let's face it - it's a dead format that needs to go the way of the 8-track tape.

It's rated for 25,000 hours (insert sound of bacon frying)

"Proprietary fanless aluminum design for superior thermal dissipation"

Then....why is it only rated for 25k hours? Don't we already have MR-16's in the 40watt equivelant range already?
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
Then....why is it only rated for 25k hours? Don't we already have MR-16's in the 40watt equivelant range already?
We'll have to see how their offerings do. Look at it this way:

They have 13W LED units rated to survive in 45C ambient. Assuming the fixture is miraculously kept at 45C, not far above room temp...

Assume 1/3 thermal efficiency; 1/3 of the energy escapes as light. The unit has to dissipate 8W of heat to a 45C environment. With a fanless aluminum design relying on no moving air, what heatsink surface temperature will be needed to maintain 8W dissipation? I see figures around 15 C/W. 45C + 125C! That's ridiculously, dangerously hot. So either they have good heatsinks (10 C/W to air?) or lower power than stated.

Heck, let's assume that their light units max out at the "Burns within a few seconds" temperature of about 75C. With a deltaT of only 30C, they'd better have some revolutionary heatsinks, or understated power.
 

LEDninja

Flashlight Enthusiast
Joined
Jun 15, 2005
Messages
4,896
Location
Hamilton Canada
perhaps their new technology will trickle down to the flashlight world. Interesting article though.
"The light source in Eric Kim's hands is just 6 millimeters across"
The XP-E is 0.9 mm across.
The XM-L is 2 mm across.

That LED will make one w-i-d-e f-l-o-o-d flashlight.
Ummm...
That wide an LED will be directional. Maybe an aspheric lens mod.
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
"The light source in Eric Kim's hands is just 6 millimeters across"
The XP-E is 0.9 mm across.
The XM-L is 2 mm across.

That LED will make one w-i-d-e f-l-o-o-d flashlight.
Ummm...
That wide an LED will be directional. Maybe an aspheric lens mod.

That'd be a monster aspheric lens. I think that a 6mm LED means that they have produced a monster array of chips. But if this is a single... I'd love to take macro shots of their bond wires. Must be a work of art in there!
 

saabluster

Flashlight Enthusiast
Joined
Oct 31, 2006
Messages
3,736
Location
Garland Tx
That'd be a monster aspheric lens. I think that a 6mm LED means that they have produced a monster array of chips. But if this is a single... I'd love to take macro shots of their bond wires. Must be a work of art in there!
From what I have read they most definitely are not using an array. They are trying to reduce system cost by using one LED. I believe the comment about a 6mm light source is referring to the package size.
 

bshanahan14rulz

Flashlight Enthusiast
Joined
Jan 29, 2009
Messages
2,819
Location
Tennessee
Soraa isn't that new... At least, the big names in the company definitely aren't. They are just recently public. Also, spec sheets of their bulbs are available on the websites now, and you can see that it is a single emitter configuration, based on the optic they used.

From what I hear, they are using non-polar GaN on a GaN substrate, as opposed to sapphire, SiC, or silicon.

From what I've read, GaN wafers are close to 100 times more expensive than Sapphire wafers. That's reason enough to pull every trick in the book to boost flux/area.

While, as we all know, most companies are rushing to get the most lumens without worrying about approximating a point source (far be it!), Soraa's main goal is to get more light per wafer of raw materials. This means they are doing exactly what we want: higher radiant flux per area. They will not achieve their company goal if they are using emitters that are 6mm across. The wafers they use are only 1cm across*, so that would mean that they are wasting a large chunk of wafer to make ONE chip. I doubt that.

SaabLuster, you ready for this shiz? ;-)

*thought I read that somewhere and I can't find the reference to it anymore.

also, http://news.cnet.com/8301-11386_3-57372704-76/startup-soraa-lights-up-with-led-2.0/ <has a good picture of GaN on Sapphire vs GaN on GaN
 
Last edited:

blasterman

Flashlight Enthusiast
Joined
Jul 17, 2008
Messages
1,802
From what I'm reading that's the case.

The size of the light source has little bearing on light projection unless you are trying to force the light into the tightest possible circle with optics, which is is irrelevant here. A bare Bridgelux C1202, which is a huge light source, doesn't appear much different than a bare Rebel a few feet away and you wouldn't be able to tell the difference except for a close examination of the shadows with millimeter rule.

Just thinking, but I thought this was the perfect application for remote phosphors.

If you increase the size of the emitter surface you gain current and thermal lattitude, correct? My problem though is there are a lot of compaies doing this, and the press release just doesn't seem very innovative and is oriented towards pacifying investors. "25,000" hours is a clear and obvious flag for me that they didn't solve the thermal problem - not even close. I'm also skeptical of his 'one year payback model'. I want to see the math on that one.
 

AnAppleSnail

Flashlight Enthusiast
Joined
Aug 21, 2009
Messages
4,200
Location
South Hill, VA
I'm also skeptical of his 'one year payback model'. I want to see the math on that one.
Call it $0.10 per kWh. Compared to an 'equivalent' 50W halogen ("Uses 25% the electricity of a 50W halogen) we're saving 37.5W. So every 26.67 hours costs 1 kWh = $0.10 of electricity. Saving $0.10 per day per replacement isn't bad, really.

24/7 use = 8736 hours of use in 1 year (24*7*52) / 26.67 = $327 saved
16/6 use = 4992 hours = $187 saved.

Check my math?
 

bshanahan14rulz

Flashlight Enthusiast
Joined
Jan 29, 2009
Messages
2,819
Location
Tennessee
and you wouldn't be able to tell the difference except for a close examination of the shadows with millimeter rule.
I think what they are getting at is instead of making bulbs with multiple emitters, and thus, multiple optics, creating the strange triple shadow, they are attempting the same sort of output with a single source for more normal looking shadows.

Just thinking, but I thought this was the perfect application for remote phosphors.
Who knows. All we can see in the photos is the heatsink and the optic. Remote Phosphors could be like the Philips bulb where the phosphors are impregnated into the outer plastic shell, or it could be that the phosphors are layered in the dome of the emitter. I'd imagine both techniques could be called Remote Phosphor.

If you increase the size of the emitter surface you gain current and thermal lattitude, correct? My problem though is there are a lot of compaies doing this
That's why Soraa is going the opposite route of all the other companies. Instead of increasing surface area for cheap gains, they are completely changing the substrate-GaN interface to improve efficacy. If all we can do is hope this company succeeds, then let's hope really hard! They will be our new favorite maker of emitters if they achieve what they are shooting for.
 

slebans

Enlightened
Joined
Mar 1, 2010
Messages
457
Location
Moncton, NB Canada
Call it $0.10 per kWh. Compared to an 'equivalent' 50W halogen ("Uses 25% the electricity of a 50W halogen) we're saving 37.5W. So every 26.67 hours costs 1 kWh = $0.10 of electricity. Saving $0.10 per day per replacement isn't bad, really.

24/7 use = 8736 hours of use in 1 year (24*7*52) / 26.67 = $327 saved
16/6 use = 4992 hours = $187 saved.

Check my math?

50 watt Halogen
50 watts x 24 hours/day x 365 days = 438,000 watt hours per year
438,000 /1000 = 438 kWh
438 kWh x $.10 = $43.80

12 watt LED
12 watts x 24 hours/day x 365 days = 105,120 watt hours per year
105,120 /1000 = 105.12 kWh
105.12 kWh x $.10 = $10.51

$43.80 - $10.51 = $33.29

You would save $33.29 in electrical costs. Additionally, there would be a savings in the cost of replacement MR-16 halogen bulbs. I have seen life values of between 1500 and 3000 hours. Let's use 2000 hours at a $3 cost.
8760 hours per year / 2000 hours = 4.23 replacement bulbs
4.23 x $3.00 = $12.69

Total savings = $12.69 + $33.29 = $45.98

In summary, as long as the LED bulb costs less than $46 it will pay for itself in the first year.

Stephen Lebans
 

slebans

Enlightened
Joined
Mar 1, 2010
Messages
457
Location
Moncton, NB Canada
"25,000" hours is a clear and obvious flag for me that they didn't solve the thermal problem - not even close. I'm also skeptical of his 'one year payback model'. I want to see the math on that one.

The 25,000 hour claim has to do with meeting the minimum Energy Star specification for the 25K operational hours - L70 lumen maintenance value and a minimum 3 year warranty. A minimum of 6000 hours of testing is required so I would guess the bulb is currently under review.

I posted the math for a one year payback calculation a little further down in this thread.

Finally, I do not like this product because they are not posting the lumens emitted. I also find that 12 watts seems high in relation to competitors current offerings(Philips Energy Star cert.) and public Reference Designs that specify/work at 10 watts or less.

Stephen Lebans
 

slebans

Enlightened
Joined
Mar 1, 2010
Messages
457
Location
Moncton, NB Canada
"The light source in Eric Kim's hands is just 6 millimeters across"
The XP-E is 0.9 mm across.
The XM-L is 2 mm across.

That LED will make one w-i-d-e f-l-o-o-d flashlight.
Ummm...
That wide an LED will be directional. Maybe an aspheric lens mod.

Here is a photo of the LED - strange looking package!
http://www.ledsmagazine.com/news/9/2/11/SILHBLED02092012

Some further info on their tech:
http://www.ledsmagazine.com/news/9/2/11
Soraa, a stealthy start-up, chose Strategies in Light to make a first public appearance. Eric Kim explained the company's first product is an MR16 LED lamp designed to replace a 50W halogen. However, the company's core innovation is to develop GaN-on-GaN LEDs. Because the LED material has a 1000-fold reduction in dislocation density, the LEDs can be driven much harder (250 A/cm2) than traditional LEDs, while suffering much less droop (the drop-off in efficiency at higher current densities). There are a whole range of ways in which the GaN-on-GaN LEDs are different from more conventional devices, not least the triangular chips (see photo). But a number of observers were sceptical due to the high cost of the GaN starting material. Look for a more detailed review of Soraa's technology and products in LEDs Magazine in the near future.

Stephen Lebans
 

Mr. Tone

Flashlight Enthusiast
Joined
Jul 2, 2009
Messages
2,350
Location
Illinois
What exactly were we looking at in that pic? This all seems interesting, though. If they really are going for higher surface brightness or not I guess we will see soon enough. Let's hope our guys can get some samples soon to put all of our speculation to rest.
 

Kinnza

Enlightened
Joined
Nov 15, 2005
Messages
223
Location
Spain
What exactly were we looking at in that pic? This all seems interesting, though. If they really are going for higher surface brightness or not I guess we will see soon enough. Let's hope our guys can get some samples soon to put all of our speculation to rest.

It's a pic of the chip used. As you can see, triangular shape, not rectangular as usual. That allows for a better yield from a circular wafer. But the main point GaN on GaN allows to use larger current densities (250A/cm2 vs <150A/cm2 for GaN on saphire or SiC) suffering of less droop. Seems interesting for throwers, as it result on higher surface luminance. They have little advantage at low-medium current levels, because GaN on GaN is not as refined yet, but offer an interesting alternative to overdriven chips. On average, GaN on GaN allows to use about double current for a given size chip with less droop on efficiency.

As always, high current densities requires of a very good thermal path. I believe high current densities only worth on a few special applications, but for sure flashlights is one of them. But Soraa will need to be competitive on larger markets to suceed and be able to offer LED packages for us. Maybe we should write them and suggest we may be excellent costumers for them. 1.5- 2 sq mm chips would blow out XML and SSTs

One note for the hope is Sunitomo and Soitec has just announced their success manufacturing 4 and 6 inch GaN wafers. Such breakthrough could allows for cheaper production of GaN on GaN chips.
 

SemiMan

Banned
Joined
Jan 13, 2005
Messages
3,899
It's a pic of the chip used. As you can see, triangular shape, not rectangular as usual. That allows for a better yield from a circular wafer. But the main point GaN on GaN allows to use larger current densities (250A/cm2 vs <150A/cm2 for GaN on saphire or SiC) suffering of less droop. Seems interesting for throwers, as it result on higher surface luminance. They have little advantage at low-medium current levels, because GaN on GaN is not as refined yet, but offer an interesting alternative to overdriven chips. On average, GaN on GaN allows to use about double current for a given size chip with less droop on efficiency.

As always, high current densities requires of a very good thermal path. I believe high current densities only worth on a few special applications, but for sure flashlights is one of them. But Soraa will need to be competitive on larger markets to suceed and be able to offer LED packages for us. Maybe we should write them and suggest we may be excellent costumers for them. 1.5- 2 sq mm chips would blow out XML and SSTs

One note for the hope is Sunitomo and Soitec has just announced their success manufacturing 4 and 6 inch GaN wafers. Such breakthrough could allows for cheaper production of GaN on GaN chips.


Do phosphors hold up at super high intensities? There would certainly be more issues with point source heating of the phosphors leading to other losses.

I wonder if they can keep up with GaN on Silicon when cost and efficiency come into play. It's great if you can be twice as bright, but if you are 3 times the cost there is no advantage.
 
Top