LED when will it become good enough?

[Soren M]

Hello,

when will the LED be good enough for us light fixture/luminare and
lightdesigners ?

when can i use LED in my desklamp and with good enough light properties (i
think it is called colour rendering) so i can work with reel colours in that
light and trust what i see ?

when will the problem with heat be solved ?

when will they be cheap enough ?

when will the Power LED´s or other LED´s be as good and energyefficient as a
CFL ?

Regards
Søren M
www.momsendesign.com

 

[Victor Roberts]

Hello,

when will the LED be good enough for us light fixture/luminare and
lightdesigners ?

when can i use LED in my desklamp and with good enough light properties (i
think it is called colour rendering) so i can work with reel colours in that
light and trust what i see ?

when will the problem with heat be solved ?

when will they be cheap enough ?

when will the Power LED´s or other LED´s be as good and energyefficient as a
CFL ?

Regards
Søren M
www.momsendesign.com

Interesting question. I attended a major LED conference in February
2000. The LED scientists at that meeting predicted that white LEDs
would have an efficacy of 30 lm/W by 2002 and 50 lm/W by 2005. It is
now early 2004 and the most efficient white LED I can find has a
guaranteed efficacy of only about 10 lm/W. (Please note I am using an
efficacy value that the LED manufacturer is willing to guarantee - not
the "typical" value or the "best in the laboratory" value.) So, in
early 2004 we have not yet met the efficacy predictions for 2002. And,
these LEDs are still very expensive on a per lumen basis.

 

[AC/DCdude17]

X-No-Archive: Yes

Hello,

when will the LED be good enough for us light fixture/luminare and
lightdesigners ?

I'll consider LEDs when all of these are met:

one unit can produce ~2600 lumens

CRI=80+

95% lumen maintenance at 40% the rated life

one 2,600 lumen unit doesn't cost more than $5 for 20,000 hours of
useful time (so, if it can last 40,000 hours and maintains 95% after
16,000 hours, then it's acceptable to me if it costs $10 each)

efficacy of lamp driver+lamp unit must be greater than or equal to 85
lumens per watt.

 

[Andrew Gabriel]

X-No-Archive: Yes




I'll consider LEDs when all of these are met:

LED's don't have to be the best lights available in order to
be adopted, but they do need some features which they can do
better than other light sources, and they don't yet have enough
such features, which is why they are currently only a minority
curiosity in lighting. They are currently running from behind
in features, and it remains to be seen if they can catch up with
existing lighting technologies, or if something new will appear
and overtake them before they get off the ground.

Just like the two men faced with a hungry lion; to live they don't
have to be able to outrun the lion, they just have to be able to
outrun the other man.

one unit can produce ~2600 lumens

That's ridiculous -- there are plenty of lamps in use with much
lower outputs than that. In my house, I have one at 180 lumens,
and a few at 350 lumens. Actually, I have only one light in the
whole house which is as much as 2600 lumens.

CRI=80+

Colour temperature of 2700K before I'd use them in the home so
they blend with other lighting and are the right colour for home
lighting levels. Probably 3500K for use in an office. All the
ones I've seen so far are much too high colour temperature for
general purpose lighting.

95% lumen maintenance at 40% the rated life

one 2,600 lumen unit doesn't cost more than $5 for 20,000 hours of
useful time (so, if it can last 40,000 hours and maintains 95% after
16,000 hours, then it's acceptable to me if it costs $10 each)

That's just as silly as your first assertion.
A 350 lumen compact fluorescent retrofit is about $4 in the UK
for ~8000 hours. An LED retrofit would need to be similarly priced
if people are going to buy it. (Selling at twice the price based on
twice the lamp life would probably significantly reduce the market.)
A 2,600 lumen lamp is probably not what any designer should start
considering for LED technology -- they are just way off on all fronts.

efficacy of lamp driver+lamp unit must be greater than or equal to 85
lumens per watt.

In terms of efficiency, they don't need to be that high to be viable.
20 lumens/watt would get them into the market. However, they need to
solve the heat destruction problem -- one way is to become much more
efficient, but there may be other solutions. If you can make a 20W
equivalent light output MR16 which doesn't cook the LEDs, that would
be a good start. I think technology only manages around 5W currently.

Other features of LEDs which might make them viable today in some
applications would be any that require the common red/yellow/green
colours with narrow beam and a tiny light source -- for this LEDs
are pretty impossible to beat at low lumen levels already.

 

[Clive Mitchell]

LED's don't have to be the best lights available in order to be
adopted, but they do need some features which they can do better than
other light sources, and they don't yet have enough such features,
which is why they are currently only a minority curiosity in lighting.
They are currently running from behind in features, and it remains to
be seen if they can catch up with existing lighting technologies, or if
something new will appear and overtake them before they get off the ground.

They've already got the edge with colour mixing, since they don't colour
shift excessively at low levels like tungsten, and they do low levels
better than cold cathode. The same great colour output at low
intensities makes them well suited to low level lighting applications.

The torch market is beginning to see LED domination in many areas.

 

[AC/DCdude17]

X-No-Archive: Yes

LED's don't have to be the best lights available in order to
be adopted, but they do need some features which they can do
better than other light sources, and they don't yet have enough
such features, which is why they are currently only a minority
curiosity in lighting. They are currently running from behind
in features, and it remains to be seen if they can catch up with
existing lighting technologies, or if something new will appear
and overtake them before they get off the ground.

Just like the two men faced with a hungry lion; to live they don't
have to be able to outrun the lion, they just have to be able to
outrun the other man.




That's ridiculous -- there are plenty of lamps in use with much
lower outputs than that. In my house, I have one at 180 lumens,
and a few at 350 lumens. Actually, I have only one light in the
whole house which is as much as 2600 lumens.

I was talking about what it would take to convince me to use it for
general purpose lighting in practical sense beyond curiosity. In
*general purpose* lighting, it's usually prefereble to have the high
output spread over length like current fluorescent lamps.

The usage pattern and power consumption of your little 180 and 350 lumen
lamps are not great enough that moving to LED wouldn't do diddly squat.

approx 2600 per lamp is quite the norm for 4' lamps and they are very
common for general purpose lighting.

Colour temperature of 2700K before I'd use them in the home so
they blend with other lighting and are the right colour for home
lighting levels. Probably 3500K for use in an office. All the
ones I've seen so far are much too high colour temperature for
general purpose lighting.

To each his own on here. I like 4100K and 5000K.

That's just as silly as your first assertion.
A 350 lumen compact fluorescent retrofit is about $4 in the UK
for ~8000 hours. An LED retrofit would need to be similarly priced
if people are going to buy it. (Selling at twice the price based on
twice the lamp life would probably significantly reduce the market.)
A 2,600 lumen lamp is probably not what any designer should start
considering for LED technology -- they are just way off on all fronts.

How many watts is the said 350 lm CFL? I've seen some miniature 5 to
10W CFLs, but they're exotic. CFLs with a four digit lumen is far more
common than such little ones.

Most white LEDs are HORRIBLE at lumen maintenance. Something around 50%
after 6,000 hrs. The only one I know of that is close enough to my
expectation is Luxeon white LED which supposedly maintains 90% after
9,000 hrs. *2002 Lumileds Lighting
(http://www.lumileds.com/pdfs/protected/AB07.PDF)

In terms of efficiency, they don't need to be that high to be viable.
20 lumens/watt would get them into the market. However, they need to
solve the heat destruction problem -- one way is to become much more
efficient, but there may be other solutions. If you can make a 20W
equivalent light output MR16 which doesn't cook the LEDs, that would
be a good start. I think technology only manages around 5W currently.

Other features of LEDs which might make them viable today in some
applications would be any that require the common red/yellow/green
colours with narrow beam and a tiny light source -- for this LEDs
are pretty impossible to beat at low lumen levels already.

LEDs have their place in spelialty lighting. I'm just saying I won't
consider them as general purpose lighting until all of the criteria I've
listed are met or comes reasonbly close to them.

The general purpose lamps (Sylvania 32W) I've got here are 48" long,
makes 3,000 lumens fresh from factory, 2,850 lumens after 8,000 hrs, CRI
in 80s and has a mean life of 20,000 hrs. All for about $3.50 a tube.

A couple of 20 lumen white LEDs would easily cost more than $3.50 even
though you need over a hundred of them to meet 2,600 lumen output.

 

[Don Klipstein]

They've already got the edge with colour mixing, since they don't colour
shift excessively at low levels like tungsten, and they do low levels
better than cold cathode. The same great colour output at low
intensities makes them well suited to low level lighting applications.

The torch market is beginning to see LED domination in many areas.

I do see how LEDs have made their incremental gains that have been
achieved to some extent or another in above-mentioned markets.

I advise a bit of "further reading" in:

http://www.misty.com/lede.html

But for now, I see most (maybe all) manufacturer/model/applications
of LEDs underperforming both hot cathode fluoreswcent lamps and
incandescent lamps both on luminous-efficacy-basis and bottom-line-basis,

With the exception of some LED-favorable applications that I mention in
http://www.misty.com/~don/lede.html (or in pages that I link therefrom)

- Don Klipstein ([email protected])

 

[Scott]

Just last night I attended a talk by Osram's chief semi-conductor lighting
engineer, who made it clear that it will be 10-20 years before we will see
any widespread use of LED technology as a substitute for
tungsten/fluorescent lamps in the domestic and commercial markets.

That is a direct opinion of a man who is at the cutting edge of the
technology, undiluted by salesman's rhetoric. At present we are struggling
to reliably get 50 lumens/Watt (reliably is the keyword here), and he
predicts it will take 5 years to achieve 100 lumens per Watt. The barrier
then is what he calls the cost:benefit ratio, which needs to become the same
as or better than that for tungsten/fluorescent sources.

Those in the know realise that the limitations of the current technology are
efficacy, economy and reliability. It will take time for those areas to be
addressed, and until that occurs, LEDs are still seen as "niche" market
sources.

Regards
Scott Forbes

 

[Andrew Gabriel]

I was talking about what it would take to convince me to use it for
general purpose lighting in practical sense beyond curiosity. In
*general purpose* lighting, it's usually prefereble to have the high
output spread over length like current fluorescent lamps.

General purpose lighting isn't always high lighting levels.
LED's could probably start looking at the lower lighting level
applications long before they become viable (if ever) for the
higher lighting level applications.

The usage pattern and power consumption of your little 180 and 350 lumen
lamps are not great enough that moving to LED wouldn't do diddly squat.

Precisely. This is the area which they are nearest to being able
to tackle, but they are still a long way from being viable even
for this.

To each his own on here. I like 4100K and 5000K.

These require significantly significantly higher lighting levels
than you will normally find in residential home lighting. White
LEDs are going to be twice as efficient at 5000K than at 2700K
(at least for the fluorescent component of the light output),
but you are going to need very much more than twice the lighting
level to avoid the cold blue appearance of high colour temperature.

How many watts is the said 350 lm CFL? I've seen some miniature 5 to
10W CFLs, but they're exotic. CFLs with a four digit lumen is far more
common than such little ones.

The 350 lm is 7W and 180 lm is 4W.
The 4W one lights a decorative lamp, which looks like a large
white glass egg with a couple of legs. I had a 7W lamp in it
originally, but it was too bright.

Most white LEDs are HORRIBLE at lumen maintenance. Something around 50%
after 6,000 hrs. The only one I know of that is close enough to my
expectation is Luxeon white LED which supposedly maintains 90% after
9,000 hrs. *2002 Lumileds Lighting
(http://www.lumileds.com/pdfs/protected/AB07.PDF)

The ones I've played with also have poor colour mixing, with
different parts of the beam containing different amounts of blue
light in particular. (No idea if these were lumileds though.)

 

[Ian Stirling]

Just last night I attended a talk by Osram's chief semi-conductor lighting
engineer, who made it clear that it will be 10-20 years before we will see
any widespread use of LED technology as a substitute for
tungsten/fluorescent lamps in the domestic and commercial markets.

That is a direct opinion of a man who is at the cutting edge of the
technology, undiluted by salesman's rhetoric. At present we are struggling
to reliably get 50 lumens/Watt (reliably is the keyword here), and he
predicts it will take 5 years to achieve 100 lumens per Watt. The barrier
then is what he calls the cost:benefit ratio, which needs to become the same
as or better than that for tungsten/fluorescent sources.

He's guessing.
Personally I suspect it's going to be 10 years or so before we see 100lm/w
in the lab, and 15 till it's in products.

Let's take a fixture with a 36W lamp that is replaced every 10000 hours,
in use 8h/day, and assume that the fixtures last the same time, with
both the same efficiancy.

Fixture initial cost $30.

360KWh = $36.
Say the cost of bulb replacement is $10 total.
So the total operating cost per year is some $13/year.

For the LED solution to be more expensive for bulk lighting, assuming an investment
rate of 5%, if the initial cost is over around some $60 per lamp, if
the LED lamp lasts forever.

This is a significant improvement over current LEDs, which
get some 4lm/W or so, 15 fold improvement would get there.

But it only needs to be a little worse than fluorescant fixtures
in electricity consumption (say 75lm/W) and any savings you get from
not changing bulbs in most applications (unless this is very expensive
and cannot be done in a planned way) go away.

 

[TKM]

Just last night I attended a talk by Osram's chief semi-conductor lighting
engineer, who made it clear that it will be 10-20 years before we will see
any widespread use of LED technology as a substitute for
tungsten/fluorescent lamps in the domestic and commercial markets.

That is a direct opinion of a man who is at the cutting edge of the
technology, undiluted by salesman's rhetoric. At present we are struggling
to reliably get 50 lumens/Watt (reliably is the keyword here), and he
predicts it will take 5 years to achieve 100 lumens per Watt. The barrier
then is what he calls the cost:benefit ratio, which needs to become the same
as or better than that for tungsten/fluorescent sources.

Those in the know realise that the limitations of the current technology are
efficacy, economy and reliability. It will take time for those areas to be
addressed, and until that occurs, LEDs are still seen as "niche" market
sources.

Regards
Scott Forbes

The US DoE, at a meeting yesterday reviewing the state of LED development,
is looking toward 2015 for practical, economical white-light LEDs with good
color quality. They were showing 30 lpw devices, but there are lots of
problems to solve. Predicting anything beyond 3-5 years is, of course,
guess work; but who needs LEDs to replace general lighting when there are so
many intereresting things to do with them right now.

Terry McGowan

 

[Don Klipstein]

This is a significant improvement over current LEDs, which
get some 4lm/W or so,

Are these white LEDs? Do they achieve 41 lumens/watt "when used as
directed", at a usual characterization current that is a majority of
absolute maximum continuous current? Can you name manufacturer, part
number, etc., and anything else I need to obtain them? Where I can get a
datasheet supporting such figures?

Preferably not relying on some premium rank that is not available to
everyone and that is some small minority of current production of that
part number.

- Don Klipstein ([email protected])

 

[Ian Stirling]

Are these white LEDs? Do they achieve 41 lumens/watt "when used as
directed", at a usual characterization current that is a majority of
absolute maximum continuous current? Can you name manufacturer, part
number, etc., and anything else I need to obtain them? Where I can get a
datasheet supporting such figures?

Err, that's not a 1.

 

[Don Klipstein]

Err, that's not a 1.

Oops! (It was late after a big workday)

Meanwhile, according to the mfr datasheets, white "Luxeon III" by
Lumileds has an overall luminous efficacy at 700 mA of typically 25.1 and
minimum 19.2 lumens/watt - at a chip temperature of 25 degrees C. I would
advise looking at that curve in the "DS46" datasheet for light output as a
function of temperature, however, since I do not expect an internal
temperature that low to be a real-world condition.
Nichia's website mentions a couple that, at 350 mA, have a luminous flux
output and voltage drop that work out to about 27 lumens/watt.

- Don Klipstein ([email protected])

 

[Ian Stirling]

Oops! (It was late after a big workday)

Meanwhile, according to the mfr datasheets, white "Luxeon III" by
Lumileds has an overall luminous efficacy at 700 mA of typically 25.1 and
minimum 19.2 lumens/watt - at a chip temperature of 25 degrees C. I would
advise looking at that curve in the "DS46" datasheet for light output as a
function of temperature, however, since I do not expect an internal
temperature that low to be a real-world condition.
Nichia's website mentions a couple that, at 350 mA, have a luminous flux
output and voltage drop that work out to about 27 lumens/watt.

Hang on...
I think I took a figure of 25lm/W, then divided 100lm/W (fluorescent)
to get an improvement needed of 4, then got interrupted, came back, stuck
lm/W on the end of that, and continued without being critical.

 

[Victor Roberts]

The US DoE, at a meeting yesterday reviewing the state of LED development,
is looking toward 2015 for practical, economical white-light LEDs with good
color quality. They were showing 30 lpw devices, but there are lots of
problems to solve. Predicting anything beyond 3-5 years is, of course,
guess work; but who needs LEDs to replace general lighting when there are so
many intereresting things to do with them right now.

Terry McGowan

Hi Terry,

Well, in at the big LED conference in Burlingame in 2000 they were
predicting 2010. So I guess the goal is always 10 years away

The reason that the DOE needs to be talking about general lighting is
because that is the only rational for all the money they are currently
throwing at LEDs. LEDs are great for traffic signals, signs, vehicular
lighting, and similar applications. However, none of these
applications have any significant impact on the DOE's mission of
reducing energy consumption for lighting. They can only justify their
investment in LEDs by making the case that they will replace a
substantial fraction of the incandescent and even fluorescent sources
used for general lighting in the US.

 

[[email protected]]

Came across this post.
Was I pessimistic!

What were the specific numbers relative to this cost benefit ratio he talked about? Sounds like it
could have been the typical "engineering perspective" containing only the cost of development and cost to
manufacture, not including elements of "cost of ownership" relative to the end user. Were the costs assuming
fiscal budgetary cycles or longer timeframes which are frequently used outside the USA? Lots of (subjective)
assumptions have to be made for him to make those types of predictions.

Also, (IMHO) any person working for a corporation developing products, who is making a public preserntation
that makes predictions relative to that companies existing or future products MUST include some amount of
"salesman's rhetoric" or they will find themselves in the unemployment line if they stray to far from the
"Company Party Line" position. I'd be willing to bet he didn't show the potential losses other lines of
business (tunston, fluorescent, ets.) will sustain as LED's mature their development. He likely had to have
his presentation content endorsed by the company's public relations department prior to the discussion, not to
imply that they required absolute truth, rather the content supported the company's current and did not
detract from it as well. Happens in all US corporations, all divisions (eng, marketing, sales, support,
etc.), the sales people get the "slimey salesman" label because the "noose" starts a little tighter requiring
"unatural acts" to maintain their employment, usually reviewed on a 90 day basis. Not attaining the 90 day
expectations for more than one review cycle and you better have another position lined up, 'cause you won't
have your current one much longer.....

Seen it happen time and time again, sorry to say...
[email protected]

 

[TKM]

Came across this post.
Was I pessimistic!

Somewhat. Look at www.lightingfortomorrow.com

One of the criteria for LFT winners is "saleability" and there are home
builders, contractors and showroom sales people on the judging panel to
check that aspect of the luminaires. The products you see on the winners
list are selling or will sell according to them. Of course, the residential
market has been slower to embrace LED-powered luminaires because that market
is more price sensitive; but I routinely see LED lighting now in commercial
applications -- both indoors and outdoors.

3000K, 90-100 lpw white-light LEDs are available and Cree isn't known for
"unreliability".
See: http://www.cree.com/press/press_detail.asp?i=1265814791138
and, for what it's worth:
http://www.cree.com/press/press_detail.asp?i=1265232091259

Staring next year in California and the year after that nationwide (US),
A-line incandescent lamps start becoming more efficacious or they begin to
disappear for general lighting. What isn't generally understood is that as
the standard incandescent lamps are redesigned for more lumens/watt, the
price per lamp will go up substantially. Manufacturers seem ready to meet
the first level of efficacy requirements (2011-2013), but will have major
problems reaching 45 lumens/watt which is the minimum lamp efficacy required
in California on 1/1/18. So, CFLs will likely be the low cost efficient
general lighting products and LEDs will start to look a lot more economical.

Overall, I think the Osram person was too pessimistic; but then
"widespread" can have many meanings.

Terry McGowan