Cree Display at Light Fair 2006

[Victor Roberts]

Light Fair has proven to be very disappointing. So far this
year I haven't seen anything exciting or even new. We seem
to be repeating the trend started in 2004 where the major
players greatly scale down their presence at Las Vegas while
having a larger booth in New York on alternate years. GE
has a smaller then normal booth. Philips has very small
booth, just about twice the minimum size, while I haven't
found Osram Sylvania yet. (They had just the bare minimum
size booth in 2004 with only one or two staff at any one
time.)

One interesting display was the Cree booth which was
completely lit with overhead 2 ft x 2 ft LED fixtures. (I
apologize in advance for using "English" units for
dimensions, but that is how building spaces are designed in
the US. I'll convert the final result to metric.) Each of
the fixtures had 10 rows of 12 Cree XLamp white 7090 LEDs.
That's 120 LEDs per 2 ft. x 2 ft. fixture. There were 48
fixtures in the ceiling, arranged as a 6 x 8 array, for a
total of 5760 LEDs.

The Cree rep said the LEDs sold for "about $2.00" so the
cost to a non-Cree user would be $11,520 just for the LEDs,
plus more for the drivers and fixtures.

I was told that the LEDs were operating at "1 watt" which I
took to mean they were operating at their rated current of
350 mA, which gives a typical power of perhaps 1.2 watts per
LED excluding the loss in the driver. The 5760 LEDs
therefore used 7689 watts assuming a driver efficiency of
90%.

Cree rates the white XLamp 7090 LED at 57 lumens "typical"
at 25C. While I don't believe these were operating at a
junction temperature of 25C I don't have any data from Cree
to estimate the actual operating temperature so I'm going to
use Cree's value of 57 lumens for now and ask all to
remember that they are probably hotter, The 5760 LEDs
therefore would generate 328,320 lumens if Cree's data is
correct and the chip temperature was 25C.

The 48 fixtures were mounted in a 18 ft. x 18 ft. ceiling
area. There were a few additional LEDs around the edges. I
think they were green, but being somewhat color blind they
could have been red I'm going to ignore them for the
moment.

Since the LEDs are directional all the light was directed
toward the floor. The area illuminated was somewhat larger
than the ceiling, perhaps 25 x 25 feet. If so, the
illuminance at the floor should have been 525 foot candles
or 5652 Lux. The area was bright, but not that bright.

If we assume that the illuminated area was 30 x 30 feet, the
illuminance changes to 365 foot candles or 3925 Lux. Still
pretty bright, In fact, about twice as bright as I estimate
the area was.

Based on data published by Cree, the output would be 80% of
rated if the junction temperature was 120C, so that cannot
explain the big difference between my simplistic
calculations and the estimated (but not measured)
Illuminance. Perhaps Terry or others can add their thoughts
about this data.

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.

 

[TKM]

Light Fair has proven to be very disappointing. So far this
year I haven't seen anything exciting or even new. We seem
to be repeating the trend started in 2004 where the major
players greatly scale down their presence at Las Vegas while
having a larger booth in New York on alternate years. GE
has a smaller then normal booth. Philips has very small
booth, just about twice the minimum size, while I haven't
found Osram Sylvania yet. (They had just the bare minimum
size booth in 2004 with only one or two staff at any one
time.)

One interesting display was the Cree booth which was
completely lit with overhead 2 ft x 2 ft LED fixtures. (I
apologize in advance for using "English" units for
dimensions, but that is how building spaces are designed in
the US. I'll convert the final result to metric.) Each of
the fixtures had 10 rows of 12 Cree XLamp white 7090 LEDs.
That's 120 LEDs per 2 ft. x 2 ft. fixture. There were 48
fixtures in the ceiling, arranged as a 6 x 8 array, for a
total of 5760 LEDs.

The Cree rep said the LEDs sold for "about $2.00" so the
cost to a non-Cree user would be $11,520 just for the LEDs,
plus more for the drivers and fixtures.

I was told that the LEDs were operating at "1 watt" which I
took to mean they were operating at their rated current of
350 mA, which gives a typical power of perhaps 1.2 watts per
LED excluding the loss in the driver. The 5760 LEDs
therefore used 7689 watts assuming a driver efficiency of
90%.

Cree rates the white XLamp 7090 LED at 57 lumens "typical"
at 25C. While I don't believe these were operating at a
junction temperature of 25C I don't have any data from Cree
to estimate the actual operating temperature so I'm going to
use Cree's value of 57 lumens for now and ask all to
remember that they are probably hotter, The 5760 LEDs
therefore would generate 328,320 lumens if Cree's data is
correct and the chip temperature was 25C.

The 48 fixtures were mounted in a 18 ft. x 18 ft. ceiling
area. There were a few additional LEDs around the edges. I
think they were green, but being somewhat color blind they
could have been red I'm going to ignore them for the
moment.

Since the LEDs are directional all the light was directed
toward the floor. The area illuminated was somewhat larger
than the ceiling, perhaps 25 x 25 feet. If so, the
illuminance at the floor should have been 525 foot candles
or 5652 Lux. The area was bright, but not that bright.

If we assume that the illuminated area was 30 x 30 feet, the
illuminance changes to 365 foot candles or 3925 Lux. Still
pretty bright, In fact, about twice as bright as I estimate
the area was.

Based on data published by Cree, the output would be 80% of
rated if the junction temperature was 120C, so that cannot
explain the big difference between my simplistic
calculations and the estimated (but not measured)
Illuminance. Perhaps Terry or others can add their thoughts
about this data.

I can add several observations to Vic's comments. LightFair is
significantly bigger than last year and has a much greater Asian character.
The Chinese manufacturers now have a whole section of the exhibit floor and
are increasingly showing fixtures as well as lamps, components and ballasts.
It was interesting to see one Chinese company (of several) featuring their
manufacturing process for 2D CFLs. That is a lamp invented by Thorn, bought
by GE and now increasingly manufactured by someone else.

There are (so far) two technical developments of interest. One is a simple
wall-box dimmer that the company TCP has designed for integral-ballast CFL
lamps. Substitute the dimmer for a wall switch and you can dim screw-in
CFLs smoothly down to 5% of full output. The dimmer has a sensing circuit
at the low end which shuts the lamps off if one starts to flicker. This is
a long-needed product for CFL downlights, but it should have numerous
applications. I saw it operate groups of CFLs from Osram-Sylvania, Philips,
GE and, of course, TCP. No price was quoted, but availability was announced
for October.

Another product from Aurora (a ballast mfr.) which I heard about and will
try to see today is an fully electronic HID ballast that can apparently
operate and dim both MH and HPS lamps, can be set for various wattages and
line voltages, and which is also addressable via internet protocol. It
sounds like another step toward the one-ballast-for-any-lamp approach that
the fluorescent ballast folks have pursued; but I'll guess that the lamp
wattages are limited -- at best to less than 400 watts.

Maybe I'll even manage find Vic today too.

Terry McGowan

 

[Victor Roberts]

I can add several observations to Vic's comments. LightFair is
significantly bigger than last year and has a much greater Asian character.
The Chinese manufacturers now have a whole section of the exhibit floor and
are increasingly showing fixtures as well as lamps, components and ballasts.
It was interesting to see one Chinese company (of several) featuring their
manufacturing process for 2D CFLs. That is a lamp invented by Thorn, bought
by GE and now increasingly manufactured by someone else.

I need to find that display.

There are (so far) two technical developments of interest. One is a simple
wall-box dimmer that the company TCP has designed for integral-ballast CFL
lamps. Substitute the dimmer for a wall switch and you can dim screw-in
CFLs smoothly down to 5% of full output. The dimmer has a sensing circuit
at the low end which shuts the lamps off if one starts to flicker. This is
a long-needed product for CFL downlights, but it should have numerous
applications. I saw it operate groups of CFLs from Osram-Sylvania, Philips,
GE and, of course, TCP. No price was quoted, but availability was announced
for October.

The Light Fair advertising for the TCP dimmer concerns me.
They have a big "*" after the phrase "Dims Any CFL" and that
"*" is not explained on the display. I asked the TCP rep
what unmentioned limitations the "*" referred to and he said
that non-TCP CFLs operated on the dimmer would have very
short life due to the lack of proper electrode heating when
the lamp was dimmed. This is a rational explanation, though
each design would have to be evaluated to see how the
electrode voltage varied as the dimmer position was varied.
If true, it greatly limits the range of CFL lamps that can
be used with this dimmer except for a few hours of show-off
at Light Fair. The TCP rep, of course, claims that their
CFLs are designed with a better electrode and will operate
for 10,000 hours at 5% dimming on this dimmer. I have not
checked the TCP data sheets to see if this life claim is
included on the data sheets in a manner that users would
have some recourse if the lamps did not last that long when
dimmed. Considering the electrode heating issues, I would
prefer CFLs that were designed for dimming from the start.

Another product from Aurora (a ballast mfr.) which I heard about and will
try to see today is an fully electronic HID ballast that can apparently
operate and dim both MH and HPS lamps, can be set for various wattages and
line voltages, and which is also addressable via internet protocol. It
sounds like another step toward the one-ballast-for-any-lamp approach that
the fluorescent ballast folks have pursued; but I'll guess that the lamp
wattages are limited -- at best to less than 400 watts.

I believe that Aurora claims their ballast is the ONLY
"fully electronic HID ballast" a claim I still do not
understand even after talking to their technical staff on
the phone. Their patents clearly show the use of high
frequency inductors and transformers, just as used in
virtually every other electronic HID and Fluorescent
ballast.

Maybe I'll even manage find Vic today too.

Hopefully...

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.

 

[Victor Roberts]

Light Fair has proven to be very disappointing. So far this
year I haven't seen anything exciting or even new. We seem
to be repeating the trend started in 2004 where the major
players greatly scale down their presence at Las Vegas while
having a larger booth in New York on alternate years. GE
has a smaller then normal booth. Philips has very small
booth, just about twice the minimum size, while I haven't
found Osram Sylvania yet. (They had just the bare minimum
size booth in 2004 with only one or two staff at any one
time.)

One interesting display was the Cree booth which was
completely lit with overhead 2 ft x 2 ft LED fixtures. (I
apologize in advance for using "English" units for
dimensions, but that is how building spaces are designed in
the US. I'll convert the final result to metric.) Each of
the fixtures had 10 rows of 12 Cree XLamp white 7090 LEDs.
That's 120 LEDs per 2 ft. x 2 ft. fixture. There were 48
fixtures in the ceiling, arranged as a 6 x 8 array, for a
total of 5760 LEDs.

The Cree rep said the LEDs sold for "about $2.00" so the
cost to a non-Cree user would be $11,520 just for the LEDs,
plus more for the drivers and fixtures.

I was told that the LEDs were operating at "1 watt" which I
took to mean they were operating at their rated current of
350 mA, which gives a typical power of perhaps 1.2 watts per
LED excluding the loss in the driver. The 5760 LEDs
therefore used 7689 watts assuming a driver efficiency of
90%.

Cree rates the white XLamp 7090 LED at 57 lumens "typical"
at 25C. While I don't believe these were operating at a
junction temperature of 25C I don't have any data from Cree
to estimate the actual operating temperature so I'm going to
use Cree's value of 57 lumens for now and ask all to
remember that they are probably hotter, The 5760 LEDs
therefore would generate 328,320 lumens if Cree's data is
correct and the chip temperature was 25C.

The 48 fixtures were mounted in a 18 ft. x 18 ft. ceiling
area. There were a few additional LEDs around the edges. I
think they were green, but being somewhat color blind they
could have been red I'm going to ignore them for the
moment.

Since the LEDs are directional all the light was directed
toward the floor. The area illuminated was somewhat larger
than the ceiling, perhaps 25 x 25 feet. If so, the
illuminance at the floor should have been 525 foot candles
or 5652 Lux. The area was bright, but not that bright.

If we assume that the illuminated area was 30 x 30 feet, the
illuminance changes to 365 foot candles or 3925 Lux. Still
pretty bright, In fact, about twice as bright as I estimate
the area was.

Based on data published by Cree, the output would be 80% of
rated if the junction temperature was 120C, so that cannot
explain the big difference between my simplistic
calculations and the estimated (but not measured)
Illuminance. Perhaps Terry or others can add their thoughts
about this data.

My second day at Light Fair was more rewarding than the
first. I found a few more booths and had time for a more
in-depth look at some booths I just surveyed yesterday.
While there are still no new sources, there are a number of
interesting new product variations and some interesting
applications. I'm going to post a longer report either
later tonight or, more likely, some time tomorrow or
Saturday. However, I do want to update the Cree info now.

I went back to the Cree booth and the rep there told me that
the illuminance at the floor level was 380 foot-candles.
This is consistent with my rough calculation if you assume
an illuminated space of 28 x 28 feet and a 10% loss of light
output due to the junction temperature being higher than
25C.

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.

 

[TKM]

My second day at Light Fair was more rewarding than the
first. I found a few more booths and had time for a more
in-depth look at some booths I just surveyed yesterday.
While there are still no new sources, there are a number of
interesting new product variations and some interesting
applications. I'm going to post a longer report either
later tonight or, more likely, some time tomorrow or
Saturday. However, I do want to update the Cree info now.

I went back to the Cree booth and the rep there told me that
the illuminance at the floor level was 380 foot-candles.
This is consistent with my rough calculation if you assume
an illuminated space of 28 x 28 feet and a 10% loss of light
output due to the junction temperature being higher than
25C.

The Cree installation is certainly an impressive display of lumens. I was
bothered by the high chromaticity -- very bluish atmosphere -- and the
radiant heat in the lighted area which was more than perceptible.

Terry McGowan

 

[Victor Roberts]

Thanks Vic and Terry for the updates much appreciated

Apart from the dollar per lumen of LEDs still not looking like
replacing bargain store CFLs any time soon, how did you rate the colour
of the white in the Cree booth ?

Thanks
Adam

I'm a bit red-green color blind so my opinion on color is
not very valuable, but Terry has given his and I agree - to
the extent possible

--
Vic Roberts
http://www.RobertsResearchInc.com
To reply via e-mail:
replace xxx with vdr in the Reply to: address
or use e-mail address listed at the Web site.

This information is provided for educational purposes only.
It may not be used in any publication or posted on any Web
site without written permission.

 

[TKM]

I'm a bit red-green color blind so my opinion on color is
not very valuable, but Terry has given his and I agree - to
the extent possible

I won't guess about the chromaticity of the Cree installation. It's
definitely "bluish" as I indicated previously. However, the color rendering
of the Cree installation is O.K. I guess it to be 70+ recalling that, at
high illuminance levels, object colors tend to look better than at the usual
300-500 lux normally used in commercial buildings these days.

Terry McGowan