Nichia Develops 60 Lumen Per Watt White LED

[udaman]

So I was reading the article on 'droop':

http://spectrum.ieee.org/semiconductors/optoelectronics/the-leds-dark-secret/0

Then I was reading the author's comments to jtr

http://www.compoundsemiconductor.net/csc/news-details.php?cat=news&id=37614

^About Nichia's lab LED's...which were discussed last year @this time @Photonics West.

http://spie.org/app/program/index.cfm?fuseaction=exhibitors&group=7

^I see from the vendor list that Cree doesn't have a presence, but Nichia is there again this weekend. So can we expect some more 'top secret' news in the next week or so?

Mukai initially introduced the result by stating that the theoretical efficacy limit for white LEDs is 263 lm/W. Yet, when questioned by the audience he explained that Nichia has pushed this ceiling to over 300 lm/W thanks to the introduction of a new phosphor technology...One member of the audience asked if Nichia had managed to push the wavelength of its lasers beyond 488 nm, only to be met by the terse reply “top secret”.

^300lm/w @20ma...yawn. I want to see 300lm/w @>350ma

If new phosphor technology can change the limits, then what does that say for what the 'theoretical' limits are??? And how did jtr come up; with his red/green/blue white LED that has a limit of 400lm/w @100% conversion efficiency? Can't we use near absolute zero science to obtain new physics and get >400lm/w @>350ma current...with CRI >95 ?

 

[jtr1962]

And how did jtr come up; with his red/green/blue white LED that has a limit of 400lm/w @100% conversion efficiency? Can't we use near absolute zero science to obtain new physics and get >400lm/w @>350ma current...with CRI >95 ?

See here, pages 10 - 12. The limit is a simple function of the spectrum needed to give a certain CRI and the eye's sensitivity to the wavelengths in that spectrum. The maximum possible efficiency for any light source is 683 lumens per watt, and that's for monochromatic 555 nm green. Once you add in other wavelengths, efficiency decreases. You can get around 400 lumens per watt with a CRI of 80 using RGB LEDs. If you want a CRI of 95, then the limit is around 340 lumens per watt, and you need to add a fourth color. No fancy physics is going to change these limits.

Speaking of limits, Cree reaches 208 lm/W at 350 mA. We're tantalizingly close to the limits of blue plus phosphor whites, which is roughly around 270 lm/W ( or 300 lm/W if you believe Nichia ). In any case, it's less than a 50% jump from where we are to the most it can ever be. RGB can get us further as already mentioned, but for now the efficiency of green needs to be greatly increased for that to happen.

 

[LunarMist]

Speaking of limits, Cree reaches 208 lm/W at 350 mA. We're tantalizingly close to the limits of blue plus phosphor whites, which is roughly around 270 lm/W ( or 300 lm/W if you believe Nichia ).

So how much brighter will our flashlights be when those babies reach mass production? For example how efficient is the emitter in the TK11 R2 in comparison?

 

[jtr1962]

So how much brighter will our flashlights be when those babies reach mass production? For example how efficient is the emitter in the TK11 R2 in comparison?

R2 bin is 114 to 122 lumens at 350 mA. Typical Vf at 350 mA is ~3.0 volts, so that comes to 108 to 116 lumens per watt. In other words, Cree's new lab sample is nearly twice as efficient, so expect your lights to get twice as birght ( or have twice the run time at the same brightness ). :cooler:

As for when these will get to mass production, if previous history is any indication I'd say 18 months to 2 years.

 

[udaman]

R2 bin is 114 to 122 lumens at 350 mA. Typical Vf at 350 mA is ~3.0 volts, so that comes to 108 to 116 lumens per watt. In other words, Cree's new lab sample is nearly twice as efficient, so expect your lights to get twice as birght ( or have twice the run time at the same brightness ). :cooler:

As for when these will get to mass production, if previous history is any indication I'd say 18 months to 2 years.

Uh, jtr, the S2 bins where supposed to be in mass production ...about now, when they were announced as being sampled last year.

Given how long it took Cree to go from R2 to R5 as current mass production XPG, wouldn't you say, conservatively that it will be 2yrs b4, we even get that 186lm lab prototype into production, and perhaps another year...or 2 before the 200lm+ barrier is a reality?

I also note, these lab samples seem to be getting their peak efficiency at lower and lower K ratings, why is that? Previously those high binned LED's were variably from 5000K to 8000K, yes?

I'd be happier if they had a better color rendering (though doesn't necessarily have to be high CRI) LED w/150+lm/w instead of now where they are <100lm/w, and unable to be driven at currents much above 350ma.

 

[jtr1962]

Uh, jtr, the S2 bins where supposed to be in mass production ...about now, when they were announced as being sampled last year.

Given how long it took Cree to go from R2 to R5 as current mass production XPG, wouldn't you say, conservatively that it will be 2yrs b4, we even get that 186lm lab prototype into production, and perhaps another year...or 2 before the 200lm+ barrier is a reality?

The S2 bins are coming any day now, any day..... :wink:

Seriously, it didn't take Cree as long as you think to go from R2 to R5. They could have rolled out better bins in their XR-E series long before the XP-G came out, but they decided to end that product line. It took a while to design and qualify the XP-G package which incorporated their better dice. That was most of the delay. We'll see how long before these latest lab developments hit mass production. My guess is we'll be well past 150 lm/W by the end of the year.

I also note, these lab samples seem to be getting their peak efficiency at lower and lower K ratings, why is that? Previously those high binned LED's were variably from 5000K to 8000K, yes?

No surprise here. If you look at that document I linked to last post, you can see that for any given CRI, you can get better efficiency at lower color temperatures. Lower color temperatures have more green and yellow which the eye is most sensitive to.

I'd be happier if they had a better color rendering (though doesn't necessarily have to be high CRI) LED w/150+lm/w instead of now where they are <100lm/w, and unable to be driven at currents much above 350ma.

If we can make 208 lm/W with standard CRI, then we can make 150 lm/W high-CRI. Cree may not be doing this because it's likely Nichia or SSC or Lumileds hold some key patents to higher CRI phosphors. Or perhaps Cree just doesn't see the market need for higher CRI. They seem to be going after the street lighting business in a big way. You don't need high-CRI for street lighting. In fact, if the higher CRI means lower intensity then that's a bad thing. The typical CRI of 75 for Cree's LEDs is going to be a huge improvement over the mid 20s of the current sodium vapor lights. CRI 75 or so is even good for less critical indoor lighting applications ( warehouses, parking garages, some types of stores ).

If anyone is interested, I did a fluorescent to LED night light conversion which turned out rather well.

 

[LunarMist]

I'd prefer better CRI and lower temps with flashlights. I don't care about streetlights much, but one should be able to see the faces of muggers clearly enough.

 

[udaman]

I'd prefer better CRI and lower temps with flashlights. I don't care about streetlights much, but one should be able to see the faces of muggers clearly enough.

"Better CRI" is based on incan color rendering, so a lower CRI LED, might give more accurate colors than a higher CRI, not the best method of reporting color accuracy.

If U had a hi-speed strobe mode Luminus SST-90 w/2 18650 powering it @5+A, you could blind the mugger b4 they could mug U

 

[Tea]

JTR, I am in awe of your achievement. I mean, here we are, a community comprised entirely of uber-geeks .... well, one amusing hypochondriac, one genuine foaming loonie, one modestly demented ape, and a whole lot of uber geeks ..... and then there is you with an entire body of detailed knowledge devoted to ... wait for it .... light bulbs.

This is genuinely awesome, and any SF member who wants to be thought of as more than a second-rate, merely ordinary geek, well, they will have to demonstrate to my satisfaction that they deserve to be considered on the same planet as the Master Geek himself. (That's you.)

Sorry all you other Storage Forum people. So far as geekdom goes, JTR is in another league.

By the way, I'm not pulling your chain. I'm following this thread with interest, and fully intend to replace all my CFLs with something better as soon as it becomes sensibly available. Oh, an I must remember to see if I can track down a LED bulb for my internal light in the car. Those incandescent things are rubbish.

 

[Handruin]

Sorry all you other Storage Forum people. So far as geekdom goes, JTR is in another league.

By the way, I'm not pulling your chain. I'm following this thread with interest, and fully intend to replace all my CFLs with something better as soon as it becomes sensibly available. Oh, an I must remember to see if I can track down a LED bulb for my internal light in the car. Those incandescent things are rubbish.

Very well said, Tea. JTR is in another league with his knowledge and skills and I also follow this thread to be enlightened (no pun intended) of the up and coming lighting technology.

 

[jtr1962]

Thanks Tea and Doug! I glad to be the light of everyone's lives around here. :idea: Or at least not one of the dimmer bulbs on the string. :bglaugh:

Oh, and Tea, plenty of decent LED internal lights for cars these days. Just try and find something using power LEDs as those are much more likely to last as long as claimed.

 

[LiamC]

Thanks Tea and Doug! I glad to be the light of everyone's lives around here. :idea: Or at least not one of the dimmer bulbs on the string. :bglaugh:

Oh I was going to say that you brighten everyone's day

Thank you. I'm here 'til Thursday

 

[udaman]

DHID ballasts?

Wondering what jtr has to say about the contentions that digitally regulated HID ballasts are a more cost effective alternative, with 'equal' useful life, to LED distance/area lighting. Claims of higher illumination @distance, lay persons terminology of 'greater penetration'.

http://www.accendoelectronics.com/G...logy-Case-Studies-Applications.shtml#DHIDvLED

DHID digital high frequency ballasts operate within the most advantageous 100-200 kHz light frequency (Sun = 100kHz) and maximize the energy efficiency and mean time before failure specifications of MH, CMH and HPS lamps.

Sun = 100kHz, is this marketing BS?

• Please note that these efficiencies are for the LED chip only, held at low temperature in a lab. In a lighting application, operating at higher temperature and with drive circuit losses, efficiencies are much lower.
• Independent testing of commercial LED lamps designed to replace incandescent or CFL lamps showed that average efficacy was about 31 lm/W in 2008 and tested performance ranged from 4 lm/W to 62 lm/W) compared with 100-116 lm/W of DHID lighting systems.

what is the current highest fixture efficiency for LED street lighting, not 2008? Does it match intensity @distances greater than say 10m? Or is that mktn BS too?

^Guess they are giving up the mkt for flashlights to LED's, as they claim this is where LED's are best

 

[jtr1962]

Wondering what jtr has to say about the contentions that digitally regulated HID ballasts are a more cost effective alternative, with 'equal' useful life, to LED distance/area lighting. Claims of higher illumination @distance, lay persons terminology of 'greater penetration'.

http://www.accendoelectronics.com/G...logy-Case-Studies-Applications.shtml#DHIDvLED

In 2008 when that was written LEDs weren't cost effective relative to the alternatives. Now the answer is it depends. LED street lighting is starting to get there, with smaller lighting needs actually being cost competitive in terms of total cost of ownership. The claim of "greater penetration" is simply marketing speak for the LEDs of the time not having enough output. You obviously can't penetrate far if you don't have enough lumens to begin with.

Sun = 100kHz, is this marketing BS?

Do these marketing people stay up nights thinking of drivel like this? The sun is steady, as in 0 Hz. And I'm not sure their figures of 100 to 118 lm/W are right, either. Most specs for bare HID lamps ( not including ballast losses ) aren't that high except the 400 and 1000 watt ones. Granted, the idea is the electronic ballast has near zero losses, so efficiency essentially equals that of the lamp itself. But 118 lm/W is pushing my credibility. I'd say 100 lm/W average for HID. And of course they're not giving fixture efficiency which is what really matters in the end. HID is an omnidirectional light source. If you're lucky you'll get 65% of the light where you need it, 70% on the outside. The rest gets wasted in the optics. So even using their high figure, we end up with a fixture efficiency of 0.7*118 = ~83 lm/W.

what is the current highest fixture efficiency for LED street lighting, not 2008? Does it match intensity @distances greater than say 10m? Or is that mktn BS too?

98.3 lumens per watt. And that's overall fixture efficiency. Note that this is better than the best case DHID of around 83 lm/W which I just calculated.

Also, here's Cree's X-Lamp lumen maintenance document: http://www.cree.com/products/pdf/XLampXR-E_lumen_maintenance.pdf

Most of the problems mentioned by the company pushing DHID regarding poor LED lifetime have to do with poor thermal design. Keep the LEDs cool enough at 350 mA, and you can get better than 140,000 hours. Even at 1000 mA, you can reach 75,000 hours with good thermal design. Not knocking DHID. It makes sense for some things. It's just that LED becomes more competitive each year, even for things like street lighting.

 

[udaman]

vaguely related, mostly OT (for the PC crowd complainers )

Made in IBM Labs: IBM Sets World Record by Creating High-Efficiency Solar Cell Made from Earth-Abundant Materials

Breakthrough holds potential to deliver more energy at a fraction of the cost

http://www-03.ibm.com/press/us/en/pressrelease/29404.wss

^ kind of mirrors some of the LED lab efficiency experiments. It's been nearly 1/2yr since the S2 bins had been announced in 'sample' quantities, yet nowhere to be found on the commercial level. Which kind of makes you wonder if it will be next year before the prior ~180lm/w LED's go into production, and maybe 2012 before we get the recently announced 200lm/w LED< more like 2015 ? Lotsa potential, no amazing *shipping* product.

Solar Panel Drops to $1 per Watt: Is this a Milestone or the Bottom for Silicon-Based Panels?


http://www.popularmechanics.com/science/research/4306443.html

^I'm thinking commenter #28 is correct however, about polysi being @19% efficiency versus amount of space required for panels. (and no, that wasn't me doing that rant post-comment )

Guess we'll have to wait N see .

As to LA's ambitious, largest in the nation, 140k retrofit LED streetlight project. I've only been by the single 2 block strip of earliest sampled LED streetlights in Culver City area. Don't know which company got that contract, or if the units are 100w? presumably? Looks like a square array of 10x10? IMHO it's a bad location for a test run, mostly 2-3story appt buildings where there is so much ambient light from various sources outside the appt buildings, it is difficult to judge.

But I would say compared to the lower output HPS (higher 300-400w HPS on taller lightpoles illuminate nearby business streets) in the neighborhood, LED's don't make much difference. They actually seem to be slightly more harsh in shadow productions, probably due to the noticeably blue tint?

Oddly enough there is a street in Culver City which has on the same intersection, and HPS SL, and older style Mercury Vapor cobrahead, as well as the more domed, really old incandescent SL <not a retorfited HID, the real deal incandescent! There are streets very neaby that don't have *any* lighting at all, and they chose a spot where there is so much ambient light from other sources, don't understand that.

Clinton's foundation is supporting this w/$$$ grants? There was some BS about how spill light from cities causing the desert skies to no longer have the wealth of stars? Don't know what kind of spiked Kool-aid those guys are drinking, but the LED fixtures I saw don't do *anything* to reduce spill light over existing HID's. Problem can only be solved by stop people from moving into communities in the dessert, and preventing metropolis population expansion into outlying areas...ie it's a population issue.

My thinking is that the LED fixture companies need to get higher efficiency LED's, and some way to soften the light, perhaps a frosted lens? Better color rendition or tint would help. Given the glare factor, I almost prefer old greenish white Mercury Vapour over the excessive blue tinted LED.

There is a section of Silver Lake (or was it Hollywood???) where according to the latest pdf, several blocks of LED fixtures with multiple vendor samples, color coded dots representing, of 150w versions I might go take a look at those test samples to see if they are an improvement.

 

[MaxBurn]

Just spotted this this morning, I didn't realize they were this affordible now.

http://i.engadget.com/2010/05/11/the-home-depot-takes-led-lighting-mainstream-with-20-bulbs/

 

[BingBangBop]

I'm sorry, as good as that is, I would contend it won't really take off till they produce an equivalent 100W bulb for even less money (~$10). They have to be competitive with CFL's.

My minimum bulb standard is 100W. The only time I use less are in light fixtures that have heat restrictions and now that CFL's exist heat restrictions are much less of a factor than they used to be. I have no interest in buying ~3x the number of fixtures to get the same light content I now have. I may be in the stone age, but that is my feeling and If I feel that way, so will others (I am not so unique).

 

[ddrueding]

It isn't heat restrictions for me, but size restrictions. CFL has allowed higher powered bulbs to run in fixtures, if they will fit. LED has the potential to get around this as well.

 

[Will Rickards]

I'd buy that for the basement light I leave on all the time for the cat to do her business. Well when the CFL that is there dies that is.

 

[LunarMist]

429 lumens is not so good for 9W is it? I think they have a ways to go.

You can buy a case of CFLs for $16.

 

[sechs]

I'm sorry, as good as that is, I would contend it won't really take off till they produce an equivalent 100W bulb for even less money (~$10). They have to be competitive with CFL's.

I agree, but if they've managed to lick the issues with colour and narrow beams, then they're a heck of a lot closer.

Maybe I could stick one in the fridge. It's really hard on my CFLs.

 

[Handruin]

Do you leave your fridge open for extended periods of time? I would hazard a guess the average length of time a fridge door is open and the light is powered on is for 5-7 seconds. Why bother with a CFL in a fridge or even an LED?

 

[LunarMist]

Do you leave your fridge open for extended periods of time? I would hazard a guess the average length of time a fridge door is open and the light is powered on is for 5-7 seconds. Why bother with a CFL in a fridge or even an LED?

The economics of that light are not good. It would require hundreds of years to recoup the electricity.

 

[timwhit]

I like to leave the fridge door open and stare at my options for several minutes at a time.

 

[LunarMist]

That would use more juice than the light.

 

[ddrueding]

My guess would be that, since the compressor is less than 100% efficient, more energy would be consumed cooling the light back down than was consumed by the light itself. The low-heat nature of the LED makes even more sense that way.

 

[LunarMist]

My guess would be that, since the compressor is less than 100% efficient, more energy would be consumed cooling the light back down than was consumed by the light itself. The low-heat nature of the LED makes even more sense that way.

A refrigeration unit usually moves more heat energy than is required to operate. I don't know the actual efficiency, but 100WH may move 500WH for example. Of course the heat ends up in the surrounding room, which is another issue.

 

[ddrueding]

A refrigeration unit usually moves more heat energy than is required to operate. I don't know the actual efficiency, but 100WH may move 500WH for example. Of course the heat ends up in the surrounding room, which is another issue.

My bad, I was thinking about peltier coolers.

 

[jtr1962]

A refrigeration unit usually moves more heat energy than is required to operate. I don't know the actual efficiency, but 100WH may move 500WH for example. Of course the heat ends up in the surrounding room, which is another issue.

Probably closer to 3:1 for typical household refrigerators. Also, the room is a closed system, so the net amount of heat added is only whatever the compressor uses. For example, if a 100 watt compressor moves 300 watts of heat, yes, you will have 400 watts total coming out the back ( or underneath ) the fridge. However, the fridge is colder than ambient, meaning it will be absorbing heat from the room, 300 watts worth in this case*. Net amount of room heating = 400 -300, or 100 watts.

* This of course assumes steady state temperatures. When the fridge is first plugged in and cools down from room temperature, it does indeed have a higher net heating effect on the room because it's lowering the temperature of the contents. However, this is only until it reaches its design temperature. Same thing on a smaller scale if you put in warm soda. Net amount of room heating will be slightly more than the compressor uses until the soda cools down.

 

[sechs]

Do you leave your fridge open for extended periods of time? I would hazard a guess the average length of time a fridge door is open and the light is powered on is for 5-7 seconds. Why bother with a CFL in a fridge or even an LED?

No. Then again, if I can replace a 40W incandescent with 13W CFL, why would I waste the electricity?

LEDs are a better choice since they tend to be brighter when cooler, whereas CFLs tend to be dimmer. Oh, and they use even less electricity.

 

[LunarMist]

The cost of the replacement LED is still prohibitive. CFL is a poor solution for a reefer.

 

[sechs]

I think ddrueding's computer hardware habit is cost prohibitive, but it doesn't seem to stop him from doing what we'd all like to do....

 

[ddrueding]

I think ddrueding's computer hardware habit is cost prohibitive, but it doesn't seem to stop him from doing what we'd all like to do....

The cost is only prohibitive if you don't have the money. I think Lunar was aiming for cost effective, but where is the fun in that

 

[LunarMist]

There are ways to save money and ways to enjoy spending it, but refrigerator lights are not on either list. :elephant:

 

[ddrueding]

There are ways to save money and ways to enjoy spending it, but refrigerator lights are not on either list. :elephant:

I'm not sure about that. Taking the more dreary moments in day to day life and finding a way to make them slightly less dreary can be beneficial.

 

[jtr1962]

The cost of the replacement LED is still prohibitive. CFL is a poor solution for a reefer.

Unless of course you get the replacement LED for free like I did. :mrgrn: I tested a bunch of LED bulbs for someone on CPF and got to keep them. Most weren't that great ( i.e. they used 5 mm indicator LEDs which are prone to fading ). A few are decent. The one I stuck in the fridge uses indicator-type LEDs but it won't log enough hours to fade much.

 

[Bozo]

LED refrigerator bulb.

 

[LunarMist]

LED refrigerator bulb.

That is not bright enough to illuminate a damned thing. orc:

 

[time]

Just 15 lumens and a mind bogglingly poor CRI of 32!

You'd get better results from a 1.7W bulb - oh wait, that's how much power this POS draws anyway ...

AFAIK, the most common wattage for a refrigerator is 15W, although it ranges between 5 and 25W. 15W should be good for about 100 lumens.

 

[LunarMist]

This 40W bulb is a common type in the US. :rambo: