LEDs: ultrabrite versus low-current

[Walter Harley]

There are a lot of different kinds of LED on the market these days. I'm
trying to understand some of the distinctions.

I'm seeing some called "low-current" (e.g., Fairchild HLMP-D150). They
don't seem to be any different than normal LEDs, except they're
characterized at low currents. But after all, any LED works at low
current - they're generally pretty linear over several decades, far as I can
tell from the datasheets.

So, what makes something be a "low-current" LED?

And, if I'm interested in operation as a panel indicator at 1mA or so,
should I be looking at "low-current" or at "ultra-bright" LED's, to get best
visibility?

Thanks,
-walter

 

[Michael]

There are a lot of different kinds of LED on the market these days. I'm
trying to understand some of the distinctions.

I'm seeing some called "low-current" (e.g., Fairchild HLMP-D150). They
don't seem to be any different than normal LEDs, except they're
characterized at low currents. But after all, any LED works at low
current - they're generally pretty linear over several decades, far as I can
tell from the datasheets.

So, what makes something be a "low-current" LED?

And, if I'm interested in operation as a panel indicator at 1mA or so,
should I be looking at "low-current" or at "ultra-bright" LED's, to get best
visibility?

Thanks,
-walter

I don't know, but "low current" probably means "decent brightness
without a lotta current". The other day I was reaching around for an
LED to use temporarily as a "heartbeat" indicator for a uC project I was
working on, and what came out of the junk box was a red LED made by HP
which I'd had since the early 70's. At 10 milliamps it hardly showed up
there on my brightly lit workbench. I pulled out a new red LED and
tried it with the same limiting resistor. *Much* better. In fact, I
increased the limiting resistor.

 

[default]

There are a lot of different kinds of LED on the market these days. I'm
trying to understand some of the distinctions.

I'm seeing some called "low-current" (e.g., Fairchild HLMP-D150). They
don't seem to be any different than normal LEDs, except they're
characterized at low currents. But after all, any LED works at low
current - they're generally pretty linear over several decades, far as I can
tell from the datasheets.

So, what makes something be a "low-current" LED?

And, if I'm interested in operation as a panel indicator at 1mA or so,
should I be looking at "low-current" or at "ultra-bright" LED's, to get best
visibility?

Thanks,
-walter

Perhaps it is a "marketing distincion?" A marketing gimmick with no
real performance.

Their way of saying "high efficiency?"

I use "ultra-bright" (another SUBJECTIVE distinction) LED's but I buy
ones with 2,000 mcd or better at 20 milliamps - then run them at 5
milliamps.

 

[Walter Harley]

I use "ultra-bright" (another SUBJECTIVE distinction) LED's but I buy
ones with 2,000 mcd or better at 20 milliamps - then run them at 5
milliamps.

That's what I was planning on also, but I'm wondering whether there's some
advantage to the "low-current" ones. Sure does seem like just a marketing
distinction.

 

[Don Klipstein]

There are a lot of different kinds of LED on the market these days. I'm
trying to understand some of the distinctions.

I'm seeing some called "low-current" (e.g., Fairchild HLMP-D150). They
don't seem to be any different than normal LEDs, except they're
characterized at low currents. But after all, any LED works at low
current - they're generally pretty linear over several decades, far as I
can tell from the datasheets.

So, what makes something be a "low-current" LED?

Some LEDs have high efficiency at low currents, and a few have high
efficiency only at low currents. Some high brightness and high
efficiency LEDs have high efficiency only at higher currents.

Examples of low current LEDs:

1. Red GaP (GaP is normally green), AKA GaP doped with ZnO, or red with
peak wavelength (do not confuse with dominant wavelength, a color
specification that roughly but not quite exactly means hue) of 690-700 nm.
These have been in use as far back as the mid 1970's. Efficiency
maximizes at 1 mA or less. The color turns orangish, sometimes even
yellow-orange at higher currents like 30 mA, which is normally safe for
these.

2. Original silicon carbide (without gallium nitride) blue. These appear
to me no longer in production. Efficiency of these is low, but not as bad
at lower currents. Efficiency appears to me maximized at somewhere
near or less than .5 milliamp.

3. InGaN blue, green, and white. Efficiency of "regular size" ones is
usually maximized at a few mA. Cree makes a series of smaller size chips
of InGaN chemistry that they call "low current", and have efficiency
probably maximized at 2 mA or a little less. Note that InGaN has
impressive efficiency and high brightness at "full current", even though
the efficiency is even higher at a few milliamps.

There is an older "high efficiency red" chemistry (orange-red GaAsP
on GaP substrate, related to yellow) that started seeing widespread use
sometime around 1980. Then there was GaAlAsP (starting in the mid
1980's), the first Superbright"/"Ultrabright" red. It was improved
afterwards by making the substrate transparent (GaP instead of GaAs).
Then there was InGaAsP, available in colors from slightly orangish red to
yellow-green and the ultimate chemistry of current LED technology for
these colors.
All of these along with red GaAsP (on GaAs substrate), yellow GaAsP on
GaP substrate and green GaP had lower efficiency at low currents. The
variation of efficiency with current led many people to believe that there
was a trick in human vision to take advantage of by pulsing such LEDs.
But no, human vision is surprisingly linear prior to a time-integration
process - the nonlinearity was in the LEDs.

I do suspect that a few "low current" LEDs do not have efficiency
maximized at low currents but are merely characterized for performance at
low currents. They may lose efficiency from low current less than others
of similar chemistryu. I have heard of some "low current" GaP green ones.

- Don Klipstein ([email protected], http://www.misty.com/~don/ledx.html)

 

[[email protected]]

Hi,

two remarks:

2. Original silicon carbide (without gallium nitride) blue. These
appear to me no longer in production. Efficiency of these is low, but
not as bad at lower currents. Efficiency appears to me maximized at
somewhere near or less than .5 milliamp.

I disagree: at least some of these are surprisingly efficient, and the
intensity of these appears to be a fairly linear function of current,
without any appreciably change in the whitish blue (sky-blue) color.

Out of curiosity I recently picked up a 3mm diffused SiC LED at a Conrad
store (2 Euro). The accompanying data sheet (bilingual German/English,
English worse than German) calls it a LDD 304, but doesn't mention the
manufacturer (perhaps Russian?).

2theta(1/2) is about 50 deg as specified, the maximum current according
to the datasheet is 30mA, the measured voltage drop is 2.99V at 5mA,
3.20V at 10mA, and 3.48V at 20mA. The apparent brightness clearly
exceeds that of a Kingbright L934SYC (judged by looking at them next to
each other with 20 mA passing through both, and by directing their light
onto a reflecting white surface). The non-diffused "superbright" L934SYC
(InGaAlP, about 10 lm/W) also has a 2theta(1/2) of 50 deg and is
specified 700 mcd (typ.) at 20mA. This should put the SiC-LED at 1000
mcd plus, not bad for a diffused LED, whatever its chemistry; peering at
it hurts the eye.

Can anybody confirm these observations? Any suggestions as to the
manufacture of a LDD 304? By the way, the datasheet says 3 mcd at 10mA,
which is ridiculous!

Then there was InGaAsP, available in colors from slightly orangish red
to yellow-green and the ultimate chemistry of current LED technology
for these colors.

You mean InGaAlP! (Arsenic and aluminum are in different columns of the
periodic table; the former is poisonous, the later is not.)

Martin.

 

[Walter Harley]

So, what makes something be a "low-current" LED?

Some LEDs have high efficiency at low currents, and a few have high
efficiency only at low currents. Some high brightness and high
efficiency LEDs have high efficiency only at higher currents.

Examples of low current LEDs:[...]

Thanks, Don! That's helpful.

 

[Don Klipstein]

two remarks:



I disagree: at least some of these are surprisingly efficient, and the
intensity of these appears to be a fairly linear function of current,
without any appreciably change in the whitish blue (sky-blue) color.

Out of curiosity I recently picked up a 3mm diffused SiC LED at a Conrad
store (2 Euro). The accompanying data sheet (bilingual German/English,
English worse than German) calls it a LDD 304, but doesn't mention the
manufacturer (perhaps Russian?).

2theta(1/2) is about 50 deg as specified, the maximum current according
to the datasheet is 30mA, the measured voltage drop is 2.99V at 5mA,
3.20V at 10mA, and 3.48V at 20mA. The apparent brightness clearly
exceeds that of a Kingbright L934SYC (judged by looking at them next to
each other with 20 mA passing through both, and by directing their light
onto a reflecting white surface). The non-diffused "superbright" L934SYC
(InGaAlP, about 10 lm/W) also has a 2theta(1/2) of 50 deg and is
specified 700 mcd (typ.) at 20mA. This should put the SiC-LED at 1000
mcd plus, not bad for a diffused LED, whatever its chemistry; peering at
it hurts the eye.

Can anybody confirm these observations? Any suggestions as to the
manufacture of a LDD 304? By the way, the datasheet says 3 mcd at 10mA,
which is ridiculous!

I suspect the LED was actually an older type GaN, lacking its actual
datasheet (and reference to its true nature) and sold in place of the SiC
ones when the SiC ones were all gone. Older GaN, such as Nichia's
now-obsolete NLPB series, had spectrum and color similar to those of SiC -
except more of an azure blue, not turquoisish. The broadband GaN had
performance that was rather linear for an LED, and had close to the same
color at 2 mA as at 30 mA. SiC gets greener at low currents and closer to
azure blue at high currents.

You mean InGaAlP! (Arsenic and aluminum are in different columns of the
periodic table; the former is poisonous, the later is not.)

Oops, I typoed - I did mean InGaAlP.

Martin.

- Don Klipstein ([email protected], http://www.misty.com/~don/ledx.html)

 

[[email protected]]

I suspect the LED was actually an older type GaN, lacking its actual
datasheet (and reference to its true nature) and sold in place of the
SiC ones when the SiC ones were all gone. Older GaN, such as Nichia's
now-obsolete NLPB series, had spectrum and color similar to those of
SiC - except more of an azure blue, not turquoisish. The broadband GaN
had performance that was rather linear for an LED, and had close to the
same color at 2 mA as at 30 mA. SiC gets greener at low currents and
closer to azure blue at high currents.

Thanks for the suggestion! I hadn't suspected this, but I'm getting more
and more convinced that you must be right. (My experience with blue LEDs
is rather limited so far.)

I had noticed though that the voltage drop was higher than specified in
the SiC datasheet (2.7V min., 3.0V typ., 3.2V max. @ 20mA). The measured
drop is lower than the datasheet values for a typical Nichia NSPB500
(InGaN 465nm to 470nm blue), but closely matches those of a typical
NSPG500 (InGaN 520nm green).

One thing struck me as somewhat unusual: the flange of this 3mm LED has
a flat at the 90 deg position from the plane of the lead frame (when
view from below with the (slightly shorter) cathode lead at 12 o'clock
the flat is at 3 o'clock). Might this permit to narrow down the
manufacturer?

The epoxy is quite strongly diffused, so that the chip cannot be
discerned when either off or on, and it is tinted in a light greenish
blue color, which appears to be well matched to the emitted wavelength
band.

Martin.