Searching \ for '[EE] : Lumens what are we talking about here ?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=
Search entire site for: 'Lumens what are we talking about here ?'.

Exact match. Not showing close matches.
PICList Thread
'[EE] : Lumens what are we talking about here ?'
2010\02\04@152134 by M. Adam Davis

face picon face
Lumens may mean the total power emitted from a source, or the total
landing on a particular surface.

The projector is specified in terms of light landing on the projection
screen - a lot of light is lost internally.

The LED is probably specified as total theoretical light power.

Further, Lumens specifically measures light output *as perceived by
the human eye*.  By pulsing the LED at higher than rated current, "the
human eye" may perceive significantly more light than running the LED
at its DC rating.

Keep in mind, however, that in most cases lumens are really only
comparable within each product.  Projectors use ANSI lumens, but you
can't compare them to lightbulb lumens, nor LED lumens.

Just like there are a million ways to measure gas mileage, there are a
million different models of the human eye, and therefore a million
ways to measure light output.

-Adam

On Thu, Feb 4, 2010 at 2:41 PM, YES NOPE9 <spam_OUTyesTakeThisOuTspamnope9.com> wrote:
> Cree announces a new LED product.....
> http://www.cree.com/products/xlamp_mpl.asp
>
> The LED package is rated at a max of 1500 lumens.   My LCD projector
> is rated at 700 lumens using an expensive bulb that puts out lots of
> heat. Probably a Hi-pressure Xenon bulb.
>
> What is the comparison between these two light sources ?
>
> Gus
>

2010\02\04@182807 by Chris McSweeny

picon face
On Thu, Feb 4, 2010 at 8:21 PM, M. Adam Davis <.....stienmanKILLspamspam@spam@gmail.com> wrote:
> Further, Lumens specifically measures light output *as perceived by
> the human eye*.  By pulsing the LED at higher than rated current, "the
> human eye" may perceive significantly more light than running the LED
> at its DC rating.
Actually no, it doesn't work like that. If you pulse an LED, the
perceived brightness will be the brightness if it was on the whole
time multiplied by the duty cycle. There is no magic to pulsing with
LEDs.

The real point with lumens being based on the eye's response is that
the eye is more sensitive to some frequencies than others, hence if
your light source has more power at that frequency it will have more
lumens than if the frequencies are equally spread. A pure 555nm green
source can be as much as 683lm/W, whilst an ideal white light source
is only 251lm/W.

Anyway, back to the original LED - 1500lm is good, but as mentioned
that's a spread out multi-emitter. Looking at other Cree products, you
can get as much as 800lm for a concentrated quad die, or 300lm for a
single die, either of which you can focus into a tight beam.

2010\02\04@191255 by Jake Anderson

flavicon
face
Chris McSweeny wrote:
> On Thu, Feb 4, 2010 at 8:21 PM, M. Adam Davis <stienmanspamKILLspamgmail.com> wrote:
>  
>> Further, Lumens specifically measures light output *as perceived by
>> the human eye*.  By pulsing the LED at higher than rated current, "the
>> human eye" may perceive significantly more light than running the LED
>> at its DC rating.
>>    
> Actually no, it doesn't work like that. If you pulse an LED, the
> perceived brightness will be the brightness if it was on the whole
> time multiplied by the duty cycle. There is no magic to pulsing with
> LEDs.
>  
http://www.ledsmagazine.com/news/5/5/11

2010\02\04@192209 by YES NOPE9

flavicon
face
>
> On Feb 4, 2010, at 4:28 PM, Chris McSweeny wrote:
>
> On Thu, Feb 4, 2010 at 8:21 PM, M. Adam Davis <.....stienmanKILLspamspam.....gmail.com>  
> wrote:
>> Further, Lumens specifically measures light output *as perceived by
>> the human eye*.  By pulsing the LED at higher than rated current,  
>> "the
>> human eye" may perceive significantly more light than running the LED
>> at its DC rating.
> Actually no, it doesn't work like that. If you pulse an LED, the
> perceived brightness will be the brightness if it was on the whole
> time multiplied by the duty cycle. There is no magic to pulsing with
> LEDs.
>
I believe that an LED using 1A at 50% duty cycle will look brighter to  
a human than an LED using 0.5A continuos.  Correct me if I am wrong.

{Quote hidden}

A Xenon bulb is not a focused point source.  It radiates more or less  
in a sphere and projector optics sort that out to produce a forward  
beam.  So I still wonder if a 1500 lumen CREE is viable as a projector  
light source.
Gus

2010\02\04@213527 by Daniel Serpell

picon face
Hi!

On Thu, Feb 4, 2010 at 9:12 PM, Jake Anderson <EraseMEjakespam_OUTspamTakeThisOuTvapourforge.com> wrote:
> Chris McSweeny wrote:
>> On Thu, Feb 4, 2010 at 8:21 PM, M. Adam Davis <stienmanspamspam_OUTgmail.com> wrote:
>>
>>> Further, Lumens specifically measures light output *as perceived by
>>> the human eye*.  By pulsing the LED at higher than rated current, "the
>>> human eye" may perceive significantly more light than running the LED
>>> at its DC rating.
>>>
>> Actually no, it doesn't work like that. If you pulse an LED, the
>> perceived brightness will be the brightness if it was on the whole
>> time multiplied by the duty cycle. There is no magic to pulsing with
>> LEDs.
>>
> http://www.ledsmagazine.com/news/5/5/11

Interesting research, but wrong conclusion :-) ,  you can download the paper at:
 http://www.jstage.jst.go.jp/article/jlve/32/2/32_161/_article

If you see graphs from 7 to 18, the actual perception versus power applied at
the blue and green LEDs are smaller with pulsed light.  This is because the
LEDs are a lot less efficient at higher currents.

   Daniel.

2010\02\05@174404 by Chris McSweeny

picon face
On Fri, Feb 5, 2010 at 12:12 AM, Jake Anderson <@spam@jakeKILLspamspamvapourforge.com> wrote:
> Chris McSweeny wrote:
>> On Thu, Feb 4, 2010 at 8:21 PM, M. Adam Davis <KILLspamstienmanKILLspamspamgmail.com> wrote:
>>
>>> Further, Lumens specifically measures light output *as perceived by
>>> the human eye*.  By pulsing the LED at higher than rated current, "the
>>> human eye" may perceive significantly more light than running the LED
>>> at its DC rating.
>>>
>> Actually no, it doesn't work like that. If you pulse an LED, the
>> perceived brightness will be the brightness if it was on the whole
>> time multiplied by the duty cycle. There is no magic to pulsing with
>> LEDs.
>>
> http://www.ledsmagazine.com/news/5/5/11

Well there's the obvious big flaw as pointed out by Daniel - they
suggest rather optimistically that LED device researchers should
concentrate on improving efficiency for pulsed operations - which
basically means reducing droop, something which isn't going away any
time soon, if ever. So the perceived brightness for the same power
input isn't actually better. How can a serious researcher can say
"With this method, the brightness of LED with a luminance efficiency
of 100lm/W can be simulated by using a 50lm/W LED," based on those
results?

Apart from that, a 5% duty cycle is itself quite a big flaw, as you
only get any benefit running LEDs at very low power levels, which is
totally the opposite of where this topic started with the excitement
of just how many lumens you could get out of an LED. I have LED
powered head torches and bike lights and even their dimmest battery
saving setting is more than 5% of full power. I'm also pretty
sceptical about how this very specific research will translate into
the real world where you have multiple emitters and are viewing
reflected rather than direct light.

On Fri, Feb 5, 2010 at 12:22 AM, YES NOPE9 <RemoveMEyesTakeThisOuTspamnope9.com> wrote:
> I believe that an LED using 1A at 50% duty cycle will look brighter to
> a human than an LED using 0.5A continuos.  Correct me if I am wrong.

You're wrong. At 50% duty cycle an LED will definitely look dimmer
than the same LED run at half the current and 100% duty, as this
effect doesn't happen at all.

Chris

2010\02\05@175442 by Chris McSweeny

picon face
On Fri, Feb 5, 2010 at 12:22 AM, YES NOPE9 <spamBeGoneyesspamBeGonespamnope9.com> wrote:
> A Xenon bulb is not a focused point source.  It radiates more or less
> in a sphere and projector optics sort that out to produce a forward
> beam.  So I still wonder if a 1500 lumen CREE is viable as a projector
> light source.

The fact something radiates in a sphere doesn't mean it's not a point
source - anything but. Whilst a Xenon bulb isn't a perfect point
source, it's not bad relatively - certainly a lot better than LEDs.
The whole point of it being a point source is that it does make it
easy to focus the light using very simple optics, whereas with
something bigger you'll always tend to be projecting an image of the
light source.

On Fri, Feb 5, 2010 at 7:38 AM, William "Chops" Westfield
<TakeThisOuTwestfwEraseMEspamspam_OUTmac.com> wrote:
> I dunno.  The Cree fits all those emitters onto a package of about 1cm
> on a side, so it's got a pretty high brightness density, or whatever
> you want to call it.

That many lumens in something 1cm on a side is rubbish surface
brightness. In fact looking at the datasheet it appears to have 24
dies, which are presumably standard cree 1mm square dies so the die
area is rather smaller, but that's still not good. Even the other LEDs
I mention do far better - 300lm LED die is 1mm a side, so 20% of the
lumen output on 4% of the area.

> If this goes on, I'd predict that someone comes out with an LED-based
> projector whose optics correct for the non-point-ness of the
> lightsource.  Fancy optics can pretty much be cast, these days, nearly
> as cheaply as less fancy optics?

Well there's fancy optics and fundamental laws of physics. Essentially
the larger your light source the larger optic you need to focus it.
That LED will need a huge optic to produce anything like a decently
focussed beam. It really is not at all suitable for use in a projector
- intended application is general illumination.

Chris

2010\02\08@102703 by M. Adam Davis

face picon face
On Fri, Feb 5, 2010 at 5:44 PM, Chris McSweeny <RemoveMEcpmcsweenyspamTakeThisOuTgmail.com> wrote:
> So the perceived brightness for the same power
> input isn't actually better.

Ah, earlier when I was asserting that pulsing a LED can make it appear
brighter I was not talking about efficiency.  I was talking about
significantly overdriving an LED during the pulse.

For instance, if an LED is rated for 10mA you may be able to drive it
with 2A pulses at a 5 or 10% duty cycle, and it will appear brighter
than simply running it steady state at 10mA.

In this case we don't care about energy efficiency, and I don't think
the press release mentions that it's energy efficient to drive it in
such a manner that it attains its maximum lumens.

Whether it's energy efficient or not is irrelevant.  The question for
this particular press release is how do they measure the lumens, and
under what drive conditions does the LED emit that power.

Or, in other words, a discussion on energy efficiency would be a
separate discussion altogether - while important in general, it's not
what the OP wanted to know.

-Adam

2010\02\08@161616 by Chris McSweeny

picon face
On Mon, Feb 8, 2010 at 3:27 PM, M. Adam Davis <stienmanEraseMEspam.....gmail.com> wrote:
> For instance, if an LED is rated for 10mA you may be able to drive it
> with 2A pulses at a 5 or 10% duty cycle, and it will appear brighter
> than simply running it steady state at 10mA.

Not for very long you won't. See
http://www.gardasoft.com/pdf/APP930%20Overdriving%20LEDs.pdf - the
recommendation is overdriving at no more than 10 times for 5% duty
cycle, or 5 times for 10% duty cycle. They don't give a figure for the
required duty cycle for 200 times overdriving (my suspicion is that it
would last very little time if you even switch on only for long enough
for it to ramp up to full power with 0.1% duty!) Note that none of the
figures there allow you to drive the LED with higher mean current than
it is rated for, they are all significantly less, in order to keep the
LED temperature down. It has to be significantly lower than the steady
state mean since the junction will heat up whilst it is being driven,
so the peak temperature will be higher than the mean. As discussed
before, for a given mean current any LED will appear brighter if that
current is constant rather than PWM. You simply can't make an LED
brighter than by driving it with constant current at its maximum rated
current.

> Whether it's energy efficient or not is irrelevant.  The question for
> this particular press release is how do they measure the lumens, and
> under what drive conditions does the LED emit that power.

As I understand it, the standard lab testing method is using a very
short pulse at the rated current and measuring the output using an
integrating sphere. Note that it's a single very short pulse, the
point of which is to keep the junction temperature at the ambient
measured temperature, not a series of short PWM pulses. Obviously not
particularly useful in the real world, but that's how they measure.
Out in the real world hobbyists seem to use a very large heatsink to
keep the temperature down when testing, but in that case you still
have the temperature gradient between the junction and the heatsink
pad - strangely no figures for that LED cluster on the datasheet,
though the single die LED I've used most (which actually has an
industry leading thermal resistance!) you get a ~30K difference
between the two when driving at full power.

Chris

2010\02\08@163350 by M. Adam Davis

face picon face
On Mon, Feb 8, 2010 at 4:16 PM, Chris McSweeny <EraseMEcpmcsweenyspamgmail.com> wrote:
> for a given mean current any LED will appear brighter if that current is constant rather than PWM.

So it sounds like all those people who are pulsing their LEDs for
increased brightness at a given power input are doing so for no
reason?

> You simply can't make an LED brighter than by driving it with constant current at its maximum rated current.

Interesting!  I've heard differently, but of course cannot easily find
out where, so I may have simply misunderstood.

I suppose I'm going to have to do some experiments and find out what
my own visual perception is.

Too bad Russell is in China, I bet he'd have a field day on this topic...!

2010\02\08@164209 by Dario Greggio

face picon face
M. Adam Davis ha scritto:
> On Mon, Feb 8, 2010 at 4:16 PM, Chris McSweeny <RemoveMEcpmcsweenyEraseMEspamEraseMEgmail.com> wrote:
>> for a given mean current any LED will appear brighter if that current is constant rather than PWM.
>
> So it sounds like all those people who are pulsing their LEDs for
> increased brightness at a given power input are doing so for no
> reason?

Well, IMO you can achieve "brighter/stronger pulses of light", useful
for instance in remote controls. But the average light would be almost
the same, and actually I don't remember using "overdrive" (let's call it
so) in other uses...


--

Ciao, Dario
--
Cyberdyne

2010\02\08@165222 by Spehro Pefhany

picon face
At 04:33 PM 08/02/2010, you wrote:
>On Mon, Feb 8, 2010 at 4:16 PM, Chris McSweeny <RemoveMEcpmcsweenyspam_OUTspamKILLspamgmail.com> wrote:
> > for a given mean current any LED will appear brighter if that
> current is constant rather than PWM.
>
>So it sounds like all those people who are pulsing their LEDs for
>increased brightness at a given power input are doing so for no
>reason?
>
> > You simply can't make an LED brighter than by driving it with
> constant current at its maximum rated current.
>
>Interesting!  I've heard differently, but of course cannot easily find
>out where, so I may have simply misunderstood.

Early LEDs had a kind of threshold (maybe a good part of 1mA) below
which no light came out.

Pulsing them gave significantly higher visual brightness, for low
average current. Modern LEDs are visible at uA in a dark room.

There is typically some nonlinearity to apparent brightness vs. current,
but it's not very significant. Might be 20-30%, which is barely
discernable. Some LEDs have curves that are convex, some concave,
depending on the type.

But heating goes up as I^2*R so you don't gain hardly
anything since your maximum *average* current must decrease to keep
the LED chip temperature the same if you want to do a apples-apples
comparison.

>I suppose I'm going to have to do some experiments and find out what
>my own visual perception is.

Try comparing two closely-matched LEDs (when connected in series they
look to be the same brightness at a reasonable current). One with DC
current, the other with pulsed current. Adjust the DC current until
they look to be the same brightness. Put a switch in there to swap
the LEDs to be sure. Measure the average power in each case and compare.

The exception is that if the pulsing is slower than the rate at which
the pulses visually fuse together in your eye, then you will get higher
apparent brightness at lower average power with the pulsed LED.

But then you see flashing rather than an apparently continuously lit LED.

>Too bad Russell is in China, I bet he'd have a field day on this topic...!

Probably. ;-)

2010\02\08@165737 by Herbert Graf

picon face
On Mon, 2010-02-08 at 17:00 -0500, Spehro Pefhany wrote:
> Early LEDs had a kind of threshold (maybe a good part of 1mA) below
> which no light came out.
>
> Pulsing them gave significantly higher visual brightness, for low
> average current. Modern LEDs are visible at uA in a dark room.

I wonder, this "pulsed is brighter" thing is something I've heard about
alot in relation to IR LEDs, does any of this hold true when using IR
LEDs?

TTYL

2010\02\08@171004 by Richard Prosser

picon face
It's not just brightness however. A flashing light is generally more
visible/noticable than a steady one. I don't know what frequency this
takes effect. Probably <20Hz I would imagine.

RP

On 9 February 2010 09:41, Dario Greggio <RemoveMEadpm.toTakeThisOuTspamspaminwind.it> wrote:
{Quote hidden}

> -

2010\02\08@172245 by Chris McSweeny

picon face
On Mon, Feb 8, 2010 at 10:10 PM, Richard Prosser <RemoveMErhprosserKILLspamspamgmail.com> wrote:
> It's not just brightness however. A flashing light is generally more
> visible/noticable than a steady one. I don't know what frequency this
> takes effect. Probably <20Hz I would imagine.

Depends of course what you're using an LED for - the high output LEDs
we're talking about here are white ones which tend to be used for
illumination - visibly flashing lights aren't particularly useful for
this!

Chris

2010\02\09@094827 by Michael Rigby-Jones

flavicon
face


> -----Original Message-----
> From: piclist-bouncesSTOPspamspamspam_OUTmit.edu [spamBeGonepiclist-bouncesSTOPspamspamEraseMEmit.edu] On
Behalf
{Quote hidden}

I realise these number were somewhat arbitrary, but I suspect they are
way off the mark. An increase from 10mA to 2A would increase the I^2
losses by a factor of 40,000, coupled with the IV losses which would be
more than 200 times higher since Vf would be significantly increase
being overdriven by that amount.

To keep the average power levels within safe levels, the duty cycle is
likely to be very small in this case, much lower than 5% and probably a
tiny fraction of 1%.

The efficiency discussion is completely relevant in this case, because
driving an LED harder does not give you proportionally more light output
above some threshold.  As you increase peak current and reduce duty
cycle to maintain safe power dissipation you will actually be getting
less average light out of the device.  As others have mentioned, for
frequencies above the persistence of vision, the eye tends to integrate
light rather than respond to the peak so it's the average light output
you are interested in.  Hence pulsing the LED is not likely to give an
improvement in brightness.

Regards

Mike

=======================================================================
This e-mail is intended for the person it is addressed to only. The
information contained in it may be confidential and/or protected by
law. If you are not the intended recipient of this message, you must
not make any use of this information, or copy or show it to any
person. Please contact us immediately to tell us that you have
received this e-mail, and return the original to us. Any use,
forwarding, printing or copying of this message is strictly prohibited.
No part of this message can be considered a request for goods or
services.
=======================================================================

2010\02\09@120638 by Peter

picon face
I believe that the discussion "pulsed is brighter than continuous" does not
refer to "pulsed fast enough to appear as if continuously on" LEDs. Maybe some
authors misquote things out of context.

Human brains and eyes are very good at seeing single events in contrast, for
example a led flashing for only 100usec once a second can be seen fairly well in
a normally lit room even if the pulse current is only 5mA. That's equivalent to
a duty factor of 1/10000 or to a ridiculously low average LED current of 0.5uA,
visible in a normally lit room! Having the LED run at 0.5uA DC under the same
conditions would likely cause the viewer not to be able to tell whether it is
on, even in a darkened room.

I could recently prove myself this fact with a USB to RS232 adapter which has a
TX LED. The cable appeared to be dead due to interface incompatibility but then
i opened a terminal and sent spaces or other characters and could see the LED
blip once per character, very faintly. I later realized that the baud rate was
256kBauds or higher and that the LED had been on for a ridiculously short amount
of time. I confirmed this switching to 300Bauds and there the LED made fat wide
pulses.

Some smoke alarms and the like use the pulse-led-once-in-X-seconds method as a
'heartbeat' and battery life monitor. The average current consumed for this is
very low and affordable. For example pulsing a LED on with 5mA every 10 seconds
for 1msec would lead to the same 1:10000 duty cycle or 0.5uA average current
consumption. This 0.5uA is low enough to be used for a clock that is meant to
run for a year on a single cell, and for similar purposes.

So, I think it's more about contrast in slow pulsing and human ability to detect
it, and not about fast pulsing at all. Also, there are special slow pulsing
frequencies which are particularly well noticed. I am not sure if they are in
the same range with those which cause seizures and must be avoided.

 Peter


2010\02\09@134432 by YES NOPE9

flavicon
face
{Quote hidden}

Were you able to discern color when looking at the 256kBaud blip ?    
Is it a rods and cones issue in the human eye ?
Gus

2010\02\09@135909 by William \Chops\ Westfield

face picon face

> visual perception

I dunno.  Isn't the usual manifestation that multiplexed and/or  
charlieplexed LED displays don't seem nearly as dim as they ought to  
be, given their actual on-time.  I've seen 3 or 4 digit displays and  
done 20-led dot-matricies that light one segment or LED at a time, and  
they certainly don't look ~1/20th the of the "normal rated current"  
brightness of the individual LEDs...

BillW

2010\02\09@141932 by Isaac Marino Bavaresco

flavicon
face
Em 9/2/2010 16:59, William "Chops" Westfield escreveu:
>  
>> visual perception
>>    
> I dunno.  Isn't the usual manifestation that multiplexed and/or  
> charlieplexed LED displays don't seem nearly as dim as they ought to  
> be, given their actual on-time.  I've seen 3 or 4 digit displays and  
> done 20-led dot-matricies that light one segment or LED at a time, and  
> they certainly don't look ~1/20th the of the "normal rated current"  
> brightness of the individual LEDs...
>
> BillW
>  


Perhaps that's because most human senses aren't linear, but rather
logarithmic such as hearing.
This may be the case of sight also.


Regards,

Isaac
__________________________________________________
Faça ligações para outros computadores com o novo Yahoo! Messenger
http://br.beta.messenger.yahoo.com/

2010\02\09@162809 by Barry Gershenfeld

picon face
On Mon, Feb 8, 2010 at 1:56 PM, Herbert Graf <EraseMEhkgrafspamEraseMEgmail.com> wrote:

>
> I wonder, this "pulsed is brighter" thing is something I've heard about
> alot in relation to IR LEDs, does any of this hold true when using IR
> LEDs?
>

In careful trials with pulsed IR LEDs and steady powered ones, participants
reported the two brightness levels were exactly the same.  They couldn't see
either of them.

I'm attempting to be funny, but Dario also mentioned remote controls, and we
want to keep in mind that we're discussing perceived brightness by the eye,
and that the effect on electronic sensors is going to be a different
discussion.

2010\02\09@171655 by Dario Greggio

face picon face
Barry Gershenfeld ha scritto:
> In careful trials with pulsed IR LEDs and steady powered ones, participants
> reported the two brightness levels were exactly the same.  They couldn't see
> either of them.
>
> I'm attempting to be funny, but Dario also mentioned remote controls, and we[...]

:)


--

Ciao, Dario
--
Cyberdyne

2010\02\09@175945 by Chris McSweeny

picon face
On Tue, Feb 9, 2010 at 5:06 PM, Peter <@spam@plpeter2006@spam@spamspam_OUTyahoo.com> wrote:
> I believe that the discussion "pulsed is brighter than continuous" does not
> refer to "pulsed fast enough to appear as if continuously on" LEDs. Maybe some
> authors misquote things out of context.

Well maybe it does if you want to go off down that road, however
that's taking this thread completely away from the original subject of
how many lumens you can get out of an LED for continuous illumination
(which is the whole point of the LED which originally started this
thread - I don't think anybody will be using it pulsed in ultra-low
power applications!) Of course no reason why threads on here can't
drift, but certainly the original arguments about pulsing LEDs were
suggesting it might be a way of producing more light for continuous
illumination (where flicker is imperceptible).

Chris

2010\02\09@180848 by Chris McSweeny

picon face
On Tue, Feb 9, 2010 at 7:19 PM, Isaac Marino Bavaresco
<spamBeGoneisaacbavarescospamKILLspamyahoo.com.br> wrote:
> Em 9/2/2010 16:59, William "Chops" Westfield escreveu:
>>
>>> visual perception
>>>
>> I dunno.  Isn't the usual manifestation that multiplexed and/or
>> charlieplexed LED displays don't seem nearly as dim as they ought to
>> be, given their actual on-time.  I've seen 3 or 4 digit displays and
>> done 20-led dot-matricies that light one segment or LED at a time, and
>> they certainly don't look ~1/20th the of the "normal rated current"
>> brightness of the individual LEDs...
>>
>> BillW
>>
>
>
> Perhaps that's because most human senses aren't linear, but rather
> logarithmic such as hearing.
> This may be the case of sight also.

It is indeed. If you ran those LEDs with the same average power, but a
steady current they would appear brighter.

Chris

2010\02\22@065154 by Russell McMahon

face picon face

I see there has been a vast amount of fun on this thread while I was away.
Internet in China was excellent at all the locations I was at (3 "hotels"
from -1* to 5* (probably), 2 factories at one location in 1 and several in
another. All seemed to produce as good or better results as my about 5 Mbps
on a good day downhill "full speed" link here. But I largely didn't look at
PICList - far too dangerous when days never have enough hours - not that
they have here either :-). But, to LEDs.

The following is my understanding based on experience and lost and lots and
lots of data sheet wading and stuff from recovered memory from decades. No
guarantees -

- Some people used to claim a superior eye response to pulsed output of same
average power level as DC. Others claimed otherwise. Doesn't seem to be
vastly significant. Phosphor decay time in modern white LEDs can't help.

- Early LEDs and still IR ones probably had a high allowable peak to mean
power level and current ratio. So pulsing at low duty cycle high power was
allowable. Bond wire melting may have been a limit.

- Modern white phosphor LEDs almost always (ie all I'm aware of) have a low
peak max to max operating current ratio. eg a 20 MA LED may have 30 mA max
pulsed rating. A 350 mA LED MAY have a 500 mA max pulsed rating etc. These
are usually NOT absolute failure ratings, but they are very liable to be
degradation related. LED degradation occurs due to both temperature and
current independently. Long learned papers discuss why. lumileds provides
some excellent engineering papers ion this demonstrating lifetime effects if
you hold one constant and vary the other. (Not impossible but clever).
So pulsing LEDs at or beyond their peak rating is very likely to impact
their lifetimes.

- Output decreases with die temperature. White LEDs (and blue, which they
are really inside) decrease by about 30% AFAIR from 25c to max die temp.
other are somewhat better or stunningly worse. Some (Red AFAIR but maybe
not) drop to about 30% of room temperature at max die temperature.
SO pulsing may do something to mean die temperature which just may help in
some situations. eg if you had an impressively low lead frame thermal
resistance and a very very good heat sink then you MIGHT be able to persuade
the die to cool down during pulsing at the end of the pulse to lower than
the mean temperature under DC. BOTE figuring as i type without using an
envelope suggests that the arithmetic doesn't work. I think it's a dream. I
know of some people who have patented a method which is said to achieve a
slight efficiency gain but I haven't yet seen the patent documents AND i
think it unlikely.

- LED lumens per mA DECREASE slightly or more with increasing mA (ie the mA
X axis / lumen Y axis graph drops from a straight line with increasing mA. I
have seen one recently released top class white LED where this curve is
almost ruler straight 0 but I suspect they were being lazy as nothing esle I
have seen achieves this.

- Vf (forward voltage) increases with current. This means that even if
lumens  were linear against current then lumen?Watt will decrease with
increasing current. Add in the drop against current and it gets worse still.

SO peak lumens does occur at max allowed current. BUT lumen/Watt improves
with DECREASING current. for above reasons.

All the above suggests that at least for modern white phosphor LEDs, pulsing
will decrease lumen/Watt.

But, as Carl Sagan was wont to say, "I may be wrong". He never actually
thought he was, I think, and in this case I don't think I am, but I may be.
Honestly :-).

_____________

As a real word example, my "favorite" LED of recent times is the superb
Nichia Raijin aka NSPWR70CSS-K1. Not to be confused with the poor relation
'Raikoh" NZPWR70CSS which despite the close to identical name has a
different lead frame and lens and probably die. The Raikoh is by most
standards to date an excellent LED. The Raijin is utterly superb.

For 1 to 2 years now it has been THE most efficient commercially available
white LED in its class (<= 50 mA). Now at last new higher power LEDs are
almost equalling it at higher currents and bettering it when they are run
at a small fraction of their rated output.

The Raijin started as a 30 mA LED with a 50,000 hour lifetime. Nichia did
some tests and rerated it to 50 mA operation and a 15,000 hour lifetime
(subject to various temperature constraints). At 30 mA it gives 150 - 160
lumen /Watt. At 50 mA it's down to about 120 lumen/Watt. Vf at 50 mA is
around 3V - which is very low by most standards and helps the efficiency.
Vf spread is stunningly well controlled. If you ask nicely Nichia will
provide production sampling Vf spread graphs and the spread is superbly
small. (Vast % inside +/- 0.05V.) They also publish luminance distribution
graphs which give you some idea of eg how many high output LEDs you may hope
to find in a large mixed sample. Not enough ;-).

I have tried Raijins at 100 mA and above "just for fun". I have not tried to
measure degradation with time. It would be interesting to do so. Output is
"very bright" :-).

For comparison - a top luminance R5 bin Cree XP-G will now give over 132
lumen/Watt at around 1 Watt (350 mA) - and at full rated current of 1.5A (!)
is still over 90 lumen/Watt (from memory).

Output drops linearly from 25C to  150 C junction temperature from 100% to
72% (datasheet precision :-) ).

Importantly, forward voltage is 2.75V typical at 100 mA, 3V at 350 mA, 3.35V
at 1A, and 3.55V at 1500 mA.
IF the lumen curve was flat with mA (and it's not) that would mean that l/W
would drop to 2.75/3.55 = 77% of the 100 mA level when run at 1500 mA.
OR conversely, you get 30% more l/w at 100 mA than you do at 1500 mA.

lumen per mA 1500 mA compared to at 350 mA is 12% lower.

Combine the Vf increase from 3 to 3.55 and the lumen drop per mA by 12% and
you get a reduction to
              3/3.55 x 78% = 66%

ie lumen/Watt at 1500 mA is only 2/3 of the rating at 350 mA.

All of which suggests that pulsing this LED is liable to give you less l/W
and not more unless you do something very clever indeed.

Lifetime: Interestingly, Cree say:

Cree currently recommends a maximum drive current of 1000 mA for XLamp XP-G
white in designs seeking the ENERGY STAR* 35,000 hour lifetime rating (≥
94.1% luminous flux @ 6000 hours) or 25,000-hour lifetime rating (≥ 91.8%
luminous flux @ 6000 hours).
Please read the XLamp Long-Term Lumen Maintenance application note for more
details on Cree’s lumen maintenance testing and forecasting. Please read the
XLamp Thermal Management application note for details on how thermal design,
ambient temperature, and drive current affect the LED junction temperature.

As someone noted here recently, Cree announced > 200 l/W at 350 mA in a lab
product recently so good things are on their way.


    Russell

2010\02\22@123520 by Alan B. Pearce

face picon face
maybe a perusal of this article is in order.

http://www.edn.com/article/CA6718478.html

2010\02\22@153547 by YES NOPE9

flavicon
face
{Quote hidden}

Rzzzzzzzzzz
Do you happen to know if high brightness LEDs would benefit from  
cooling using some kind of refrigerator scheme or use of a TEC  
( thermal electric cooler ) ?
Gus

More... (looser matching)
- Last day of these posts
- In 2010 , 2011 only
- Today
- New search...