Searching \ for '[EE]: LED pulse or constant current' 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/displays.htm?key=led
Search entire site for: 'LED pulse or constant current'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: LED pulse or constant current'
2002\09\24@024802 by Wouter van Ooijen

face picon face
On some other fora I see arguments pro and con pulsing of a LED to
achieve better (perceived?) brightness with the same average current.
Does anyone know of hard (experimental) data about this?

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\09\24@064940 by Jinx

face picon face
> On some other fora I see arguments pro and con pulsing of a LED
> to achieve better (perceived?) brightness with the same average
> current. Does anyone know of hard (experimental) data about this?

Quotes from previous discussions - seems very subjective

===================================================

one of the threads in that hazily-recalled discussion was
persistence of vision and whether pulse width / frequency could be
optimised to take best advantage of it. Although brightness of LEDs
can be measured empirically, persistence of vision from person to
person is subjective, leading one to set up for "average" intensities
that are acceptable to as wide a range of viewers as possible

===============================================

Pulsed light is claimed to have a greater apparent brightness for the
same average energy. I guess the human eye acts as a peak detector.

===============================================

Much closer to an average detector. The apparent brightness only
changes slightly (usually improves) with shorter duty cycles and the same
average current. Losses increase, of course, as Vf increases.
With early (1970's) LED materials, there was a threshold of perhaps
1mA below which not much light was emitted, but with modern materials
it is much more zero-based. Super-bright GaAlAs LEDs are usable and
fairly linear over a very wide range, from 10's of microamps to scores
of mA.

It's easy to test this, all you need is a scope and a source of
square waves of different duty cycles (I'd expect you to use a PIC,
of course). You can scope the voltage across the dropping resistor to
see what the actual current and duty cycle is. Pick a couple of
matched LEDs and adjust them for equal brightness with different
duty cycles.

If the pulses are slow enough that the flicker is visible, then the
results will be different, favoring the short duty cycles if they
are long enough.

There are also some interesting shifts in the spectrum of the
emitted light with current, some of the effect due to heating of
the die, and some not

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\09\24@161058 by Bernard Boudet

flavicon
face
On Tue, 24 Sep 2002 08:46:48 +0200 you wrote:

> On some other fora I see arguments pro and con pulsing of a LED to
> achieve better (perceived?) brightness with the same average current.
> Does anyone know of hard (experimental) data about this?

HP had a very comprehensive applications note on this in one of their
books (quite old now, before even the PIC was a twinkle in Microchip's
eye ;-)).

Assuming you do mean high frequency pulsing (as opposed to flashing), it
works like this:

LEDs (although essentially constant voltage) do have a non-linear
characteristic.  I.e. if you pass a higher current, the voltage across
the LED will go up to some degree.

Power = V x I.  So by pulsing with a higher current, a higher average
power dissipation can be achieved than with the average current at d.c.
More power means more light and heat.

The LED current rating is quoted as a max. d.c. value, based on *power*
dissipation, since that is what will destroy the LED.  So you might think
the LED rating should be reduced for pulsed operation, resulting in a
net benefit of zero.

*However*, the main point of the HP app note was that the important
thing is not to exceed the maximum junction temperature.  At d.c. this
is proportional to power dissipation (plus ambient temperature), but
with pulsed operation the junction gets to cool down between pulses,
also there is a certain rise time (thermal response) during the on
periods.

So following HP's own data, it's possible to run an LED above it's
equivalent d.c. power rating (which has gotta mean more light, human
perception or not).  But to do it properly you do need to know various
parameters, which would be quoted on the data sheet for that LED (e.g.
max. junction temp., thermal resistance, etc.)

OTOH, it was a loong time ago I looked at this, and all the above could
a figment of the old grey matter, even utter bollocks.  You could do
worse than search through HP's (Siemens', et al.) application notes
though.

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\09\25@110427 by Peter L. Peres

picon face
On Tue, 24 Sep 2002, Bernard Boudet wrote:

*>books (quite old now, before even the PIC was a twinkle in Microchip's
*>eye ;-)).

Before the PIC was a twinkle in Microchip's eye it made a long career with
General Instrument who first implemented this architecture (16C54 like
with mask ROM).

There is a saying about there being no new things, only new eyes. It
probably holds true outside major academic teams and inventor's houses ...

Peter

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\09\25@153916 by Jan-Erik Soderholm

face picon face
About pulsed/blinking LED's (or other lights)

As having a son who has epilepsy, I would just remind you
that the pulse/blink freq should either be rather low (a few Hz),
or aprox 80 Hz or more.

Even if you can't "see" blink in the 40-60 Hz area, people
with epilepsy might very well have problem with them.
(We have changed our TV set from a 50Hz model to a 100Hz
model, and always setup his PC monitor with a high screen
refresh rate, but a LCD would be better...)

This is also true for e.g. multiplexed 7-seg LED displays, don't
multiplex to slow to save instruction cycles.

But perhaps you was thinking of a much higher frequency.

Well, maybe a not so often thought of fact in electronics design...

Jan-Erik Svderholm

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\09\25@175634 by Bernard Boudet

flavicon
face
On Wed, 25 Sep 2002 21:37:31 +0200 you wrote:

> About pulsed/blinking LED's (or other lights)
>
> As having a son who has epilepsy, I would just remind you
> that the pulse/blink freq should either be rather low (a few Hz),
> or aprox 80 Hz or more.

> This is also true for e.g. multiplexed 7-seg LED displays, don't
> multiplex to slow to save instruction cycles.
>
> But perhaps you was thinking of a much higher frequency.

Yes, sorry, I should have mentioned the improved efficiency I was
talking about needs 1kHz+ (IIRC).

> Well, maybe a not so often thought of fact in electronics design...

Well no, perhaps not.  I just made a PIC counter with 7-seg display and
set the scan rate until it stopped flickering (for me).  I guess most
people would do the same.

At 30Hz the display did flicker, but was quite readable.  At 60Hz it is
rock steady (to my eye).  From what you say though, perhaps 120Hz would
be more sensible?

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics



'[EE]: LED pulse or constant current'
2002\10\01@163601 by Harold M Hallikainen
picon face
On Tue, 24 Sep 2002 21:09:45 +0100 Bernard Boudet
<spam_OUTlist-piclistTakeThisOuTspamFOOBAR.CLARA.CO.UK> writes:
>
> *However*, the main point of the HP app note was that the important
> thing is not to exceed the maximum junction temperature.  At d.c.
> this
> is proportional to power dissipation (plus ambient temperature), but
> with pulsed operation the junction gets to cool down between pulses,
> also there is a certain rise time (thermal response) during the on
> periods.
>
> So following HP's own data, it's possible to run an LED above it's
> equivalent d.c. power rating (which has gotta mean more light, human
> perception or not).  But to do it properly you do need to know
> various
> parameters, which would be quoted on the data sheet for that LED
> (e.g.
> max. junction temp., thermal resistance, etc.)
>

       This does seem strange to me. If, as I expect, the limiting factor is
junction temperature, it seems that pure DC would give you maximum
output. Going to pulses, the junction temperature will have "ripple." The
magnitude of the ripple (assuming a perfect heat sink) would be
proportional to peak input power and thermal resistance to the heat sink.
It would be inversely proportional to the heat capacity of the device.
With relatively high frequency pulses, the heat capacity keeps the
junction from heating or cooling much, resulting in low ripple and a
relatively stable junction temperature (similar to driving the LED with
DC, in which case there'd be no ripple). Driving with a lower frequency
would result in a larger ripple with the same average temperature, but a
higher peak temperature, which seems like it could be a problem. It is
good that you point out that increased current also causes increased
voltage across the LED, so power dissipation increases more than just in
proportion to the current.
       So, I'd expect pure DC on an LED would allow the LED to be driven with
more power because the peak and average temperatures would be the same.
       The original question on perceived brightness is interesting. I'd like
to see any research on that. I believe that as a sort of data reduction
technique, our brain tends to notice flashing lights more than steady
lights. When something is steady, we don't refresh the image in our
brain, since it's just the same data anyway. If the image is changing (as
in a flashing LED), it needs to be refreshed, so we notice the LED more
than we'd notice a steady LED. I think this is especially true with
peripheral vision, where we are especially only looking for changes. I
would think, however, that the flash rate would have to be low enough
that we perceive it as flashing. This is all just a guess, though. Seeing
some experimental results would be interesting.


Harold




FCC Rules Online at http://hallikainen.com/FccRules
Lighting control for theatre and television at http://www.dovesystems.com

Reach broadcasters, engineers, manufacturers, compliance labs, and
attorneys.
Advertise at http://www.hallikainen.com/FccRules/ .


________________________________________________________________
GET INTERNET ACCESS FROM JUNO!
Juno offers FREE or PREMIUM Internet access for less!
Join Juno today!  For your FREE software, visit:
dl.http://www.juno.com/get/web/.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\10\02@180006 by Bernard Boudet

flavicon
face
On Tue, 1 Oct 2002 13:22:58 -0700 you wrote:

> > So following HP's own data, it's possible to run an LED above it's
> > equivalent d.c. power rating (which has gotta mean more light, human
> > perception or not).  But to do it properly you do need to know
> > various
> > parameters, which would be quoted on the data sheet for that LED

>         This does seem strange to me. If, as I expect, the limiting factor is
> junction temperature, it seems that pure DC would give you maximum
> output. Going to pulses, the junction temperature will have "ripple." The
[snip]
>         So, I'd expect pure DC on an LED would allow the LED to be driven with
> more power because the peak and average temperatures would be the same.

Well, did I mention it was a long time ago ;-).  I found an app note
from Agilent which looks suspiciously like the HP data book I
remember reading.  Only it says the opposite!  
I think *now* what I remember is that it's possible to counter the
reduced brightness of pulsed operation by increasing the current above
the max dc rating.  The idea being to pulse the LED with a high current,
but for so short a time that the junction be ok, then leave it off for a
relatively longer period to let it 'cool down'.  
The increased brightness thing is actually increased *efficiency* (of
the circuit, not the LED), by operating the LED at a higher, switched,
current.  But this is only possible at high frequency with low duty
cycle:

http://literature.agilent.com/litweb/pdf/5091-9704E.pdf

So you are right, well spotted!  Thankyou.  
But hold on a minute... What's this? :

http://www.stockeryale.com/illumination/leds/literature/app001.htm

(I would trust the HP data, YMMV.)

Cheers,
-bernie.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservKILLspamspam@spam@mitvma.mit.edu with SET PICList DIGEST in the body


2002\10\02@195808 by Harold M Hallikainen

picon face
On Wed, 25 Sep 2002 17:24:51 +0300 "Peter L. Peres" <plpspamKILLspamACTCOM.CO.IL>
writes:
> On Tue, 24 Sep 2002, Bernard Boudet wrote:
>
> *>books (quite old now, before even the PIC was a twinkle in
> Microchip's
> *>eye ;-)).
>
> Before the PIC was a twinkle in Microchip's eye it made a long
> career with
> General Instrument who first implemented this architecture (16C54
> like
> with mask ROM).
>
> There is a saying about there being no new things, only new eyes. It
> probably holds true outside major academic teams and inventor's
> houses ...
>

       A discussion of PIC history from 1999 appears at
http://www.drzyzgula.org/bob/text/ht.txt . Actually, the history of the
PIC would be a nice addition to the Microchip website...

Harold


FCC Rules Online at http://hallikainen.com/FccRules
Lighting control for theatre and television at http://www.dovesystems.com

Reach broadcasters, engineers, manufacturers, compliance labs, and
attorneys.
Advertise at http://www.hallikainen.com/FccRules/ .


________________________________________________________________
GET INTERNET ACCESS FROM JUNO!
Juno offers FREE or PREMIUM Internet access for less!
Join Juno today!  For your FREE software, visit:
dl.http://www.juno.com/get/web/.

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservKILLspamspam.....mitvma.mit.edu with SET PICList DIGEST in the body


2002\10\16@104425 by Harold M Hallikainen

picon face
On Wed, 2 Oct 2002 22:49:38 +0100 Bernard Boudet
<EraseMElist-piclistspam_OUTspamTakeThisOuTFOOBAR.CLARA.CO.UK> writes:
{Quote hidden}

       This second data does not seem to conflict with the HP data. It says
"The second reason is to increase the effective brightness of the
illuminator during the pulse by using a higher pulse current than the CW
rating, since the luminance is proportional to current." Note that it
says "DURING THE PULSE." HP also says you can take advantage of higher
LED output during with pulse operation in machine vision and other
photodetector uses, where the detector is quick enough to detect the
pulse amplitude. With human vision, however, the pulse is averaged over a
longer period of time, so we observe the average brightness instead of
the peak brightness. Note also in the HP ap note that the LED has an
efficiency peak around 20mA. So, putting a higher peak power into it
while maintaining the average power results in less light output due to
the lower efficiency during the pulse. To quote HP, "It is always better
to drive an LED device with a high dc current to obtain the necessary
light output to be viewed by a human observer than to pulse drive the
LED. Using a high peak current and a low duty factor to pulse drive an
LED device produces less time average light output than by using a high
dc drive current."
       As further pointed out in the HP ap note, at low pulse frequencies,
there is increased "temperature ripple" at the junction, with the peak
temperature above the average temperature. It is the peak temperature
that must be limited, and the average is below that. With DC drive, the
average and the peak temperatures are the same, so the average power into
the LED can be higher. At higher pulse frequencies, the ripple is
reduced, allowing the average power into the LED to be increased without
exceeding the peak junction temperature. However, during the high power
pulse, LED efficiency is lower, so there is less light output and more
junction heating. So, even at high pulse frequencies, the light output is
less with pulse than with DC drive.
       However, some have suggested on this list that human perception does not
respond to average brightness, but, instead, somewhere between average
and peak. This appears to conflict with the HP application note. I'd be
interested in reading any info suggesting human perceived brightness is
higher than average power with pulsed light. At relatively low flash
frequencies, I believe a flashing LED "catches our attention" better than
a steady LED, possibly due to data reduction techniques used in our
vision. Something that is unchanged is not refreshed in our brain. At
slightly higher frequencies, on the border of perceivable flash and not
perceivable, the flash can trigger epilepsy. I know almost nothing about
epilepsy. It sort of sounds like the data rate due to refresh is getting
too high and we're getting a buffer overflow.


Harold


FCC Rules Online at http://hallikainen.com/FccRules
Lighting control for theatre and television at http://www.dovesystems.com

Reach broadcasters, engineers, manufacturers, compliance labs, and
attorneys.
Advertise at http://www.hallikainen.com/FccRules/ .


________________________________________________________________
GET INTERNET ACCESS FROM JUNO!
Juno offers FREE or PREMIUM Internet access for less!
Join Juno today!  For your FREE software, visit:
dl.http://www.juno.com/get/web/.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


More... (looser matching)
- Last day of these posts
- In 2002 , 2003 only
- Today
- New search...