LEDs could start to replace lighbulbs in 2 years

[Paul_in_Maryland]

article.

[jtr1962]

I agree with the article but I wonder at $20 each how many households will convert to LED bulbs. It took CFLs declining in price to $1 or $2 before they started to reach reasonable market penetration, and you see lots of households where people still use regular incandescents. Given that LEDs will have none of the disadvantages of CFLs such as shortened life from frequent starts, inability to run on a standard lamp dimmer, or a delay until full brightness I wonder if government will finally force the issue by just prohibiting the sale of standard incandescents. Even at $20 the LED bulb will be very cost effective given its energy efficiency/long life. It's just that it'll be a big hit on your wallet every time you decide to replace one of your incandescents with LED. On the plus side, if the LED bulb is properly designed, it'll last at least 50,000 hours (that's nearly 23 years if used 6 hours a day). For all intents and purposes this is virtually forever.

In two years LEDs should be at 100 to 120 lm/W. This means a 14 to 17 watt LED bulb can do the job of a 100 watt incandescent. This is roughly half of what a CFL now uses. Also, LED bulbs are closer to a point source so they are more likely to be accepted in things like chandeliers than CFLs.

[hank]

Moving to a white LED based on a green instead of a blue phosphor would get rid of the spike of blue light energy now in almost all white LEDs. I hope the manufacturers are paying attention to the research -- the eye doctors and companies making replacement lenses for cataract patients are well aware of the problems, but as with all such it's easier -- and better public health practice -- to avoid the problem at the manufacturing stage than filter it at the consumer end

Between the new tech that sends more photons out the useful end, and the quantum dot coatings, and the OLED and other methods, I think there's hope.

[Ken_McE]

"Right now, consumers and businesses can buy a light-emitting diode, or LED, that provides about the same level of illumination as a... conventional 60-watt lightbulb, Steven DenBaars, a professor of material science at the University of California Santa Barbara, said at the SEMI NanoForum, taking place here this week."

If one assumes 450 lumens from a 60 watt incan, what LED fixture is he talking about? That is about the output of a white 18 led Luxeon flood board, and they go for what, one or two hundred dollars apiece? Besides the LEDs you would need a built in heat sink and inverter.

/Not sure how the article differs from wishful thinking?/

[hotbeam]

From a light output point of view, it is here and available right NOW! Cost wise, it may quite a few years (possibly never) to get down to the everyday affordable price rate, ie supermarket pricing. However for commercial use, it has a surprisingly short pay-back period when all things are taken account.

[Canuke]

Originally Posted by jtr1962

I wonder if government will finally force the issue by just prohibiting the sale of standard incandescents.

I'm inclined to doubt it unless there were some huge energy crisis that people felt needed that kind of response... rather like the ol' double nickel on the freeways.

LED's have a cool factor which will likely help push them along as prices drop and efficiencies increase.

Originally Posted by jtr1962

In two years LEDs should be at 100 to 120 lm/W. This means a 14 to 17 watt LED bulb can do the job of a 100 watt incandescent. This is roughly half of what a CFL now uses. Also, LED bulbs are closer to a point source so they are more likely to be accepted in things like chandeliers than CFLs.

LED's can also be designed with simple support circuitry that gives a steady light without any RFI-generating components -- a bridge rectifier and a capacitor. No ballasts etc.

[NewBie]

Originally Posted by Ken_McE

"Right now, consumers and businesses can buy a light-emitting diode, or LED, that provides about the same level of illumination as a... conventional 60-watt lightbulb, Steven DenBaars, a professor of material science at the University of California Santa Barbara, said at the SEMI NanoForum, taking place here this week."

If one assumes 450 lumens from a 60 watt incan, what LED fixture is he talking about? That is about the output of a white 18 led Luxeon flood board, and they go for what, one or two hundred dollars apiece? Besides the LEDs you would need a built in heat sink and inverter.

/Not sure how the article differs from wishful thinking?/

You could do it with one lamina ceramics array, and it would still sleeping:
"At greater than 1,200 lumens, the Titan daylight white model is competitive with compact fluorescent bulbs...

Boasting an 800 lumen output, the Titan RGB LED light engine produces any of 16 million saturated and blended colors (including white with variable color temperature) from a single point source. "
http://www.laminaceramics.com/docs/T...168_103006.pdf

Or you can do it with 3 CREE XR-E, which would cost you in low volume, under 3 dollars each.

The converter in low volume can be done for under 5 dollars with products from companies like Supertex, and in some models with efficiencies over 90%, and if designed right, no issues with RFI like CFL ballasts/bulbs often have:
http://www.supertex.com/pdf/misc/LED_driver_matrix.pdf

Both of these would fall significantly in light bulb type volumes.

[hank]

Look at that Titan datasheet --- their LEDs have a beautiful heat exchanger. Radial fins! that looks like an old radial-cylinder aircraft engine.

The admins here have decreed keeping a clear line between flashlights and general illumination --- but I will bet in five years it will be common to take your preferred light with you, adding a battery pack for portable, and simply plug it in when you want to use it for reading or other indoor uses. And it'll be something like those infinitely adjustable light engines.

[brickbat]

Originally Posted by NewBie

You could do it with one lamina ceramics array, and it would still sleeping:
"At greater than 1,200 lumens, the Titan daylight white model is competitive with compact fluorescent bulbs...

Alright, let me just don my 'devil's advocate' hat for a minute. Here's a few problems I see in thinking a Titan Light Engine would be a good CFL replacement, as compared to, say a Sylvania 13W Dulux (CF13/E/841):

CRI: CFL 82, Titan 78 typ (62 minimum)
Initial Lumens: CFL 800, Titan 597 (at 1050 mA drive, 100C Tj)
Power used: CFL 13W, Titan 21W (at 100C Tj)
Efficacy: CFL 62 lm/W, Titan 28 lm/W (at 100C Tj)

So, I conclude the CFL is twice as efficient, and has better CRI.

No doubt, Lamina is making some advanced components - I've spent some time tinkering w/ a RGB engine, but I don't see how the argument for their use in general lighting makes economic sense. Not yet, anyway.

[NewBie]

Originally Posted by brickbat

Alright, let me just don my 'devil's advocate' hat for a minute. Here's a few problems I see in thinking a Titan Light Engine would be a good CFL replacement, as compared to, say a Sylvania 13W Dulux (CF13/E/841):

CRI: CFL 82, Titan 78 typ (62 minimum)
Initial Lumens: CFL 800, Titan 597 (at 1050 mA drive, 100C Tj)
Power used: CFL 13W, Titan 21W (at 100C Tj)
Efficacy: CFL 62 lm/W, Titan 28 lm/W (at 100C Tj)

So, I conclude the CFL is twice as efficient, and has better CRI.

No doubt, Lamina is making some advanced components - I've spent some time tinkering w/ a RGB engine, but I don't see how the argument for their use in general lighting makes economic sense. Not yet, anyway.

LED Lighting Fixtures, Inc. (LLF) today announced performance results from its most recent recessed can fixture. The fixture was tested by an independent lab as providing 700 lumens of white light at a color temperature of 2900 Kelvin, with a color rendering index (CRI) of 93. Total power supplied to the fixture from a residential voltage, AC power source was 13 watts. The fixture efficiency equaled approximately 54 lumens per watt of warm white light.
http://www.ledlightingfixtures.com/pr_021606.pdf


LED Lighting Fixtures Inc. (LLF) today announced another record in solid state fixture performance at 80 lumens per watt from its most recent recessed downlight. This product was tested by an independent lab as providing 600 lumens of warm white light at a color temperature of 3100 Kelvin, with a color
rendering index (CRI) of 92. Total power supplied to the fixture from a residential voltage AC power source was 7.5 watts.
http://www.ledlightingfixtures.com/pr_053006.pdf


LED Lighting Fixtures was founded by Neal Hunter, the same fella that founded CREE in 1987.

[brickbat]

Interesting claims. What emitters are they using? A mid-bin Cree XR-E is good for around 80 lm at about 1.2W, 67 lm/W, but that's at a Tj of 25C I'm pretty sure - how the heck do you maintain that? At Tj of 100C, you've got more like 57 lm/W.

Color me skeptical, but I'd like to see a 'real' data sheet in lieu of 'press releases'. I'm not saying it can't or won't be done, it's just that with more than a few years' experience with semiconductor data sheets, I've found it's generally a good idea to read them very carefully. As the cliche goes, ignore the entire first page, and NEVER, ever rely on a component press release if you're responsible for designing it into a product...

[NewBie]

Originally Posted by brickbat

Interesting claims. What emitters are they using? A mid-bin Cree XR-E is good for around 80 lm at about 1.2W, 67 lm/W, but that's at a Tj of 25C I'm pretty sure - how the heck do you maintain that? At Tj of 100C, you've got more like 57 lm/W.

Color me skeptical, but I'd like to see a 'real' data sheet in lieu of 'press releases'. I'm not saying it can't or won't be done, it's just that with more than a few years' experience with semiconductor data sheets, I've found it's generally a good idea to read them very carefully. As the cliche goes, ignore the entire first page, and NEVER, ever rely on a component press release if you're responsible for designing it into a product...

They are not claims, they were independent lab tested. They should hit the market early next year. This isn't just a component, but an entire finished lighting fixture.


BTW, your numbers for CFL are pretty whacked up.
I grabbed one of my GE CFL bulbs, and this is what it shows:

49807 – FLE10HT32SWCD3PK
520 lumens
Power 10W
Life: 12,000 hours
CRI: 82
http://genet.gelighting.com/LightPro...DUCTCODE=49807

This puts it at 52 lm/W

Then you need to include the significant additional losses incurred in the fixture.

Once you do all that, go back to the independent lab testing for the LLF setup, which includes the fixture, and note the 80 lm/W, and the 43,800 hour life.


Also realize that LLF is not using the CREE XLamp, but mounting the CREE dies themselves, to materals that have lower thermal resistance and customized for the fixture.

[St8kout]

There are already ads in my sailing magazines for replacing all your internal and external boat lights with more efficient LED lights to save on battery power.

[brickbat]

Originally Posted by NewBie

They are not claims, they were independent lab tested.

Well, I contend they are claims, based on testing by their chosen test lab, and contained (presumably) in a test report they have not published. Good science is presented in a way that other scientists can refute claims made. Where I work, press releases are generated by the Sales & Marketing team. Engineers write data sheets for use by other engineers.

BTW, your numbers for CFL are pretty whacked up….

No, not whacked up.

http://genet.gelighting.com/LightPro...DUCTCODE=20532

Your numbers include the ballast losses for the CFL. I purposely chose a CFL without an internal ballast since the LED to which I was comparing it had no internal ballast electronics.

… and the 43,800 hour life.

This is why we need a real data sheet. At what level of lumen depreciation was the 43,800 hour life specified? Further, what’s the lamp (LED emitter) replacement cost? THAT matters when you bring 'life' into the mix...

But let’s go back to your first claim:

You could do it with one lamina ceramics array, and it would still sleeping:
"At greater than 1,200 lumens, the Titan daylight white model is competitive with compact fluorescent bulbs...

Where is the evidence that supports this claim?

[AlexSchira]

Banning incans would be a tad extreme, but I can imagine a little media pushing towards LEDs might become an option. A few commercials here and there, a couple news blurbs on the money saved and advantages of using LEDs, even store displays to get people competing with the Joneses next store could help. Most every house I've been to still uses incan bulbs and fixtures, from the ritzy townhouses to the trailer parks. Most people still think LEDs are those things that light up on video game consoles. So far they've caught on with the flashaholics, who in turn are campers, survivalists, or just compulsive wisebuyers.

This is why I'm a bit irked that Maglite hasn't been advertising their new LED line at all, with the power behind their brand name, Joe Sixpack will finally see the light. This may lead to 'What else can these things do?', which may make household LEDs much easier to imagine.

[jtr1962]

Originally Posted by AlexSchira

Banning incans would be a tad extreme, but I can imagine a little media pushing towards LEDs might become an option.

It may be but sometimes people will continue using one thing even when something better comes along out of force of habit. Consider that we've had linear T8 fixtures with no flicker and good color rendering for at least the last 20 years. Long term the total cost of ownership is way cheaper than incandescent fixtures yet they just haven't caught on in residential settings. In fact, it seems we've taken a step backwards. Kitchens at least more often than not used to use linear tubes. Now in a lot of new houses they have those horrid high hats with incandescents. Now if incandescents just weren't an option, or perhaps were taxed to reflect their high energy use, you might actually see linear fluorescents in most homes instead as well as a greater variety of more aesthetically pleasing fixtures for these tubes. CFLs may save energy but I see them as a halfway house designed for a general public that hasn't gotten out of the light bulb mentality.

Linear fluorescents are superior in every way to CFLs. They are more efficient, give much more even lighting, can be used in fully recessed fixtures, run cooler, etc. We've been using them in every room we're in a lot for the last 25 years. The last bastion of incandescent bulbs in our house are the three very infrequently used chandeliers in the living room, dining room, and master bedroom. Once LED bulbs are available for reasonable cost those incandescents will be gone as well. If it was my house instead of my mother's I would have put linear T8 fixtures in those rooms as well. I've long hated everything about incandescents-the heat they generate (i.e.especially when burning myself changing a burnt out light bulb), their short life, their poor quality of light, the huge amount of power they use for the little light that they generate. I'm frankly amazed that these drawbacks haven't pushed more people to use CFLs and linear fluorescent.

While all the talk of LEDs is great, and I really hope they do finally displace incandescent bulbs for good in a few years, we've already had an equally good solution right under our noses for the last 20 years but have failed miserably in pushing it. At the very least we should have required T8 fixtures in every room in new construction. If it is there then people would use it.

One thing LED has going on its side beside the coolness factor is that it can ultimately be made as cheaply as incandescent light bulbs, if not cheaper. The driver electronics are trivial, the raw materials for the LEDs themselves cost pennies. The maufacturing process, which is the bulk of the cost of LEDs, will continue getting cheaper each year. I remember well when 5mm white LEDs sold for $3. Now you can get them in quantity for under ten cents. Since people more often than not go with the alternative which is cheaper up front, this bodes well for the eventual universal adoption of LEDs.

[hotbeam]

Originally Posted by brickbat

Well, I contend they are claims, based on testing by their chosen test lab

That's a bit cynical

[NewBie]

Originally Posted by brickbat

No, not whacked up.

http://genet.gelighting.com/LightPro...DUCTCODE=20532

Your numbers include the ballast losses for the CFL. I purposely chose a CFL without an internal ballast since the LED to which I was comparing it had no internal ballast electronics.

Ah, so you are not including reality, how very convient, especially when the ballasts for CFL are typically quite significant as far as loss factors. Choose a certain special LED to demonstrate how poor they are too.

On another note, I've got plenty of LEDs that have been going continously for six years now and they have not burned out, just dropped a little in intensity. What is interesting, with more agressive cooling, to note that one can extend the lumen over time quite considerably.

GE's definition of rated life:
"The rated life is a measure of the median time in hours that it takes for a light bulb to burn out. This is the point in laboratory testing at which half the test bulbs have burned out and half the test bulbs are still burning."

People often wonder why these 10,000 hour CFLs don't last at least 10,000 hours, that is why.


GE's definition of lumen maintenance:
"Lumen maintenance – The luminous flux or lumen output at a given time in the life of the lamp and expressed as a percentage of the initial luminous flux. The mean lumens are the value at 40% of rated life."

I've noticed that with only two years on the CFL bulbs (and much less time on them, how new ones are so much brighter), how quickly they dim. So I often end up replacing them before they are dead anyhow.

Another quick thing to note is how fluorescent bulbs like to be ran for long periods of time. Turning them off and on greatly shortens their life as well as lowers their lumen output. This is borne out by new testing regiments which are slowly working their way to acceptance, due to complaints about the problem.

Originally Posted by brickbat

This is why we need a real data sheet. At what level of lumen depreciation was the 43,800 hour life specified? Further, what’s the lamp (LED emitter) replacement cost? THAT matters when you bring 'life' into the mix...

Hold your horsies, it is comming.

Meanwhile, I'll point out something from one of the CREE datasheets, where the LED die is not as agressively cooled:
"Cree has accumulated operating data in excess of 28,000 hours. Using this data, long-term lumen maintenance can be projected. The method used for extrapolation is an exponential fit of the data from 1,000 hours until the end of the data set. Based on this method, XLamp LEDs are projected to have lumen maintenance of greater than 70% on average after 50,000 hours when used in accordance with published guidelines."

Originally Posted by brickbat

But let’s go back to your first claim:

Where is the evidence that supports this claim?

That is not my claim, please take it up with Lamina Ceramics.


Another item, besides agressive heatsinking, to increase lm/W that can be employed is to under-drive the die. Since the packaging of the LED is more expensive than the die, if you are mounting the dies yourself anyhow, it quickly becomes apparent to use more die, and not to drive them as hard. A side benefit, is you also extend the lifetime further.

An example, using a CREE XR-E, and what happens to the lumen per watt, during actual testing, mounted on a 2" by 3" by 0.165" piece of copper:




So, since they are using multiple dies, and mounting them themselves, it is apparent that their numbers very easily could be reality.


.

[brickbat]

Originally Posted by NewBie

Ah, so you are not including reality, how very convient

I compared a CFL to a LED. Neither had ballast losses included. You think that's somehow not realistic?

Originally Posted by NewBie

You could do it with one lamina ceramics array, and it would still sleeping:
"At greater than 1,200 lumens, the Titan daylight white model is competitive with compact fluorescent bulbs...

He said it, but he's not standing behind it...

[jtr1962]

Originally Posted by brickbat

I compared a CFL to a LED. Neither had ballast losses included. You think that's somehow not realistic?

The reason is that ballast losses for a CFL are far more than for an LED. It's easily possible to make an LED ballast which is 96 or 97% efficient. Most fluorescent ballasts are in the 85% area. A few of the better ones of the type used to drive linear tubes may get into the low 90s. The ones on many cheapie CFLs are probably only around 80% efficient.

Note that the ratings on the packages of CFLs do in fact include ballast losses. This makes it fairly easy to calculate overall system efficiency. I've found that very small 4 watt CFLs only manage 45 to 50 lm/W. Medium-size ones of about 13 to 20 watts are around 55 to 60 lm/W. Some of the larger ones can approach 70 lm/W. All these figures are when new. Figure about 20% less midway through lamp life to allow for lumen depreciation. That gives average efficiency figures in the mid 40s to mid 50s.

Now let's look at LEDs. As NewBie said, it's quite possible to get efficiencies in the 80s by mounting Cree dies on a decent substrate. The ballast will only decrease these numbers by a few percent. Lumen depreciation is around 30% by 50,000 hours so figure average lumens over the life of the LED will be about 85% of initial lumens. If you start out around 80 lm/W then you can average in the high 60s over the life of the LED. This is a good 25% to 45% better than the average wallplug efficiency of CFLs.

Note however that LEDs still can't touch linear T5 or T8 tubes yet, at least in raw wallplug efficiency. With a decent ballast and tubes you can have average efficiencies over the lamp life in excess of 80 lm/W. However, consider that linear tubes typically have fixture losses of around 15% so that brings the average efficiency into the high 60s. This about matches LEDs. For area lighting LED really has no fixture losses unless a diffuser is used.

[NewBie]

Originally Posted by brickbat

He said it, but he's not standing behind it...

Heh, you might try barking up another tree. It was Lamina Ceramics that said that...See third paragraph:
http://www.laminaceramics.com/news/052406.aspx

As I said, go take it up with Lamina Ceramics.

[brickbat]

Lamina Ceramics didn't post it here. You did. And you added the "You could do it with one lamina ceramics array, and it would still sleeping" part.

As I read their "news release" they compare the efficieny of their LED arrays to PAR30 Halogen. Their comparison to a CFL appears to be in regards to lumen output only. Big deal.

[brickbat]

Originally Posted by jtr1962

The reason is that ballast losses for a CFL are far more than for an LED. ...Most fluorescent ballasts are in the 85% area...

A 40 Watt, 24 Vdc Advance Xitanium LED driver is rated at 80% (40.8W out, 51 W in.)

http://www.ledcentral.com/Products/8...ntelliVolt.pdf

Note that the ratings on the packages of CFLs do in fact include ballast losses.

agreed, at least for internal 'screw-in' lamps.

[NewBie]

Originally Posted by brickbat

A 40 Watt, 24 Vdc Advance Xitanium LED driver is rated at 80% (40.8W out, 51 W in.)

http://www.ledcentral.com/Products/8...ntelliVolt.pdf

From your datasheet link:
Efficiency: 90% typical

Yes, it varies from device to device, but one would assume you'd use a little wisdom and pick an appropriate ballast for the application, instead of going balls to the walls:




There are plenty of examples of AC plug-in supplies that will deliver 90% efficiencies, even at full load:
http://www.supertex.com/pdf/misc/HV9910DB1.pdf

[brickbat]

Yeah, typical. I guess we'll just have to recognize that we interepret technical data differently, you and I.

I'd still like to see the 97% efficient LED driver JTR eluded to...

[jtr1962]

Originally Posted by brickbat

I'd still like to see the 97% efficient LED driver JTR eluded to...

It's easy enough to do-just decrease the switching frequency from the usual hundreds to thousands of KHz to 50 kHz or so and use MOSFETs with lower on resistance. The only drawback is that it requires larger inductors but in a residential lighting fixture there is plenty of room for that. In the push to miniaturize everything many LED ballasts operate at high frequencies but the trade-off is lower efficiency.

BTW, I'm in the process of designing a step-down ballast to drive a single LED off of 12 volts. Efficiency is currently 85% but if I were to drive two or three LEDs in series I could easily get into the low 90s. Also, if my budget was higher I could get into the low 90s even driving a single LED, and the mid or high 90s driving two or three.

Regarding the ballast you linked to, leave it to a company which designs fluorescent ballasts to make an LED ballast that large and that inefficient. A properly designed LED ballast, even a 97% efficient one, can still be much smaller than a fluorescent ballast. I'm amazed that a company can screw up making an LED ballast by that much. The problem is that they're approaching it with the preconceptions they have from years of designing fluorescent ballasts. You really need a fresh approach since LEDs are a totally different animal.

[brickbat]

OK, now I see. 97% is easy to do, but the product doesn't quite exist yet...

Hmm.

[NewBie]

Originally Posted by jtr1962

Regarding the ballast you linked to, leave it to a company which designs fluorescent ballasts to make an LED ballast that large and that inefficient. A properly designed LED ballast, even a 97% efficient one, can still be much smaller than a fluorescent ballast. I'm amazed that a company can screw up making an LED ballast by that much. The problem is that they're approaching it with the preconceptions they have from years of designing fluorescent ballasts. You really need a fresh approach since LEDs are a totally different animal.

Most of the design of the ones you picked out was done long ago, back in 2001. Philips was having them fesigned/made and putting their label on them.

The early ones were in black.

Now, sometimes they are black, and sometimes they are white, see page 4:
http://www.philips.com.my/assets/Dow...l(1)-13545.pdf

The same Xitanium is produced by Advance, and for example, look like this:
http://www.advancetransformer.com/up...E-6010-R02.pdf

Additional Xitaniums:
http://www.ledcentral.com/DisplayPro...&rid=P&appid=2
http://www.ledcentral.com/DisplayPro...&rid=P&appid=2

An all in one:
http://www.advancetransformer.com/up...E-6040-R03.pdf


One will note, they look absolutely identical to the Philips branded ones:
http://www.lighting.philips.com/gl_e...m_UK_final.pdf
http://dianaelectronic.de/faq/luxeon_nt.pdf

[NewBie]

A BRIGHT FUTURE

Projections by the Optoelectronics Industry Development Association show that lifetime ownership costs of LED lighting systems will be less than those of incandescent lamps by next year and less than those of fluorescent lamps by 2012. Market-research firm IC Insights forecasts that high-brightness LEDs, including white LEDs used for room lighting, will account for 86% of the world's $6.7 billion total revenues for LED sales by 2010.
http://www.elecdesign.com/Articles/I...eID=13982&pg=2