Thermal and low current performance of white Luxeon LED

[JohnR66]

I was looking at some of the data over on the Lumileds site. I was wondering
how die temperature affects light output.
also interested in how efficiency l/w is affected as the current is lowered.

Not looking for exact figures, just in general how the luxeon LEDs perform
considering the above.

Also, what is the typical output of a 12 volt .5 amp incandescent bulb
(likely argon gas fill) for comparison.

Thanks, John

 

[JohnR66]

Just pull up the datasheets from the web site and look for the junction
temperature vs. light output graphs to see the effect of temperature on
light output. (Fig 2a and 2b in the Luxeon I Emitter Data Sheet).
Note that the amber and red LEDs have significantly different
temperature vs output characteristics than the other colors. There is
also a graph of light output vs. current (Fig. 4a and 4b) that should
adress your question of efficiency vs. current. Although the amount of
current driven through the LED will affect the junction temperature
which will affect the light output as well, so it's not a direct
relationship. In other words lowering the current could produce more
light than you would expect from the current derating curve because you
will also improve your junction temperature. It will be highly
dependent on the heat sink, ambient temperature, and whether there is
airflow or not.



As for a comparable incandescent source, the only one I can think of
that is close to what you listed would be the W5W miniature bulb used
in automotive sidemarkers and park lamps. This runs at 5 watts (12
volts) and produces 50 lumens.

Thanks. I poked around and found the technical datasheet. I was looking in
the PDFs at the bottom of each Luxeon page before.

Things look very good if I could keep the junction temp at -20c!

I'm thinking of a bike light application. LEDs are far more efficient when
the current is reduced when compared to the yellowing, dimming incandescent
as the generater speed drops or the batteries wear down.

 

[[email protected]]

Also, what is the typical output of a 12 volt .5 amp incandescent bulb
(likely argon gas fill) for comparison.

Somewhere between 2 and 6 candlepower. Some miniature lamp data
follows. Most of these are probably argon gas fill, except for the
ones marked "halogen"; some of them are two-filament lamps and the data
is for the less bright filament.

type volts amps cp
17916 12.00 0.42 1.20

916 13.50 0.54 2.00
2397 14.00 0.48 2.00
2057 14.00 0.49 2.00
3057 14.00 0.59 2.00

7443 12.00 0.42 2.80

3496 12.00 0.66 3.00
1034 14.00 0.59 3.00
1157 same as 1034
3157 same as 1034
3357 same as 1034
2357 14.00 0.69 3.00

17171 12.00 0.42 4.00
67 13.50 0.59 4.00
68 same as 67
1095 14.00 0.51 4.00
b1007 same as 17171
204 15.00 0.44 4.00

175 14.00 0.58 5.00

89 13.00 0.58 6.00
90 same as 89
98 13.00 0.62 6.00
1016 14.00 0.59 6.00
1176 same as 1016
205 14.00 0.60 6.00

3652 12.00 0.42 6.00 halogen
70711 same as 3652

918 12.80 0.56 6.50

208 14.00 0.68 7.00

68161 12.00 0.50 10.0 halogen

Next you're probably going to ask how to turn candlepower into lumens.
For a rough estimate that is bad for reasons that other posters here can
explain better than I can, you can multiply the candlepower number by
12.57, making the 2-6 cp range something like 25-75 lumens.

Part of the problem with this estimate, I think, is that this assumes
there's no reflector - the light is going in all directions and is
falling on a sphere, which is not the way these lamps are usually used.
I have no doubt that there are other problems with it.

Matt Roberds

 

[Don Klipstein]

I was looking at some of the data over on the Lumileds site. I was wondering
how die temperature affects light output.
also interested in how efficiency l/w is affected as the current is lowered.
Not looking for exact figures, just in general how the luxeon LEDs perform
considering the above.

The datasheets for the various Luxeon products have curves showing light
output as a function of temperature and as a function of temperature.

The curves are "normalized" to 100% at "characterization current" and at
25 degree C junction temperature.
In the "electrical characteristics" table earlier in the datasheet is a
figure for thermal resistance, in degrees C per watt. MUltiply this by
power input in watts, add this to the temperature of the heatsinkable
surface of the LED, and that is the junction temperature.

In general, Luxeons of all colors will have a little more luminous
efficacy when mildly to moderately underpowered than at full power. Same
is true of most white, InGaN blue and non-yelowish green LEDs of any
brand. All LEDs whose temperature characteristics I have seen have
efficiency varying inversely to some extent or another with temperature.

Also, what is the typical output of a 12 volt .5 amp incandescent bulb
(likely argon gas fill) for comparison.

This will vary with design life expectancy and to a little extent with
other factors such as filament shape.

First order of approximation - somewhat less than half the approx. 180
lumens of the "93" 12.8V 1.04 amp lamp, and with a little less overall
luminous efficacy (lumens out per watt in) if life expectancy is the same
(750 hours) - I would pull 60 lumens out of a hat.
Although at this moment I do have the urge to mention a rather
efficacious low power incandescent lamp - Osram's HPR50, a flashlight
lamp. It is a halogen lamp with xenon as the main fill gas ingredient and
has a rated life expectancy of 25 hours, but is designed to produce 89
lumens from 4.42 watts.
I would want it noted that incandescent lamps have some "economies of
scale" where efficacy largely increases with wattage and is also usually
favored by higher design currents (thicker filaments). Because of
this, 12 volt incandescents tend to be more efficient than higher
voltage ones of the same wattage and life expectancy. Other factors
also matter - such as low voltage design (generally near and especialy
under 5 volts) resulting in very short filament length compromising
efficiency due to heat conduction through the ends of the filaments.

- Don Klipstein ([email protected])

 

[Don Klipstein]

Thanks. I poked around and found the technical datasheet. I was looking in
the PDFs at the bottom of each Luxeon page before.

Things look very good if I could keep the junction temp at -20c!

I'm thinking of a bike light application. LEDs are far more efficient when
the current is reduced when compared to the yellowing, dimming incandescent
as the generater speed drops or the batteries wear down.

I would like to add the clarification:

With mild to moderate underpowering, white LEDs make a slight gain in
lumens out per watt in, while incandescents decrease very noticeably in
this area! Any given incandescent with variation on input power (watts)
has energy efficiency varying roughly proportionately with input watts,
usually varying slightly more than input watts does!

As an example - with an incandescent, reduce input voltage by 10%. By a
"usual one-size-fits-all" rule as well as I remember it, input amps
decreases 5% (I have yet to see or hear of more than 6%), input watts
decrease about 15%, and light output is down close to 30%.

This I find to be the main reason for LED flashlights to enjoy much
longer battery runtime than incandescent ones do!

- Don Klipstein ([email protected])

 

[Don Klipstein]

<SNIP figures of MSCP, volt and amp figures for a bunch of specific
incandescent lamps>

Next you're probably going to ask how to turn candlepower into lumens.
For a rough estimate that is bad for reasons that other posters here can
explain better than I can, you can multiply the candlepower number by
12.57, making the 2-6 cp range something like 25-75 lumens.

Part of the problem with this estimate, I think, is that this assumes
there's no reflector - the light is going in all directions and is
falling on a sphere, which is not the way these lamps are usually used.
I have no doubt that there are other problems with it.

The candlepower figures that you posted are actually MSCP, or "mean
spheerical candlepower" (for at least most and probably all of the
specific lamps mentioned). That is a measure of total light output,
with one MSCP being 12.57 (specifically 4 times pi) lumens.

MSCP and "candlepower" differ from the candela, an official metric unit.
The candela roughly means "beam candlepower" for directional light
sources, although I am suspicious of the claims of many handheld
spotlights made in recent years. Candlepower is not an official unit of
beam intensity but has been in the past, while the official unit of
photometric output per solid angle in recent decades is now the candela.

- Don Klipstein ([email protected])

 

[JohnR66]

I would like to add the clarification:

With mild to moderate underpowering, white LEDs make a slight gain in
lumens out per watt in, while incandescents decrease very noticeably in
this area! Any given incandescent with variation on input power (watts)
has energy efficiency varying roughly proportionately with input watts,
usually varying slightly more than input watts does!

As an example - with an incandescent, reduce input voltage by 10%. By a
"usual one-size-fits-all" rule as well as I remember it, input amps
decreases 5% (I have yet to see or hear of more than 6%), input watts
decrease about 15%, and light output is down close to 30%.

This I find to be the main reason for LED flashlights to enjoy much
longer battery runtime than incandescent ones do!

- Don Klipstein ([email protected])

Don,
That is why I like my Luxeon flashlights so much! I get good runtime with my
3-AAA cell 1 watt Luxeon flash light and it seems brighter a typical 2 D
cell light. My only worry is the batteries are in poor shape by the time the
LED is not putting out much usable light and could leak in the light if I
don't change them in good time.
John