"Wim Lewis" wrote in message
...
In article ,
Tom Coates wrote:
I want to build some emergency lighting equipment using LEDs and 12V. A
friend told me that it is possible to get more light per Watt by pulsing
the
LED to instantaneous levels well above its average ratings.

As I understand it, this is not true --- it was true in a way for older
LEDs, but isn't true for modern LEDs. Pulsing a diode above its rated
current could get it to emit some of its light at shorter wavelengths,
which
were closer to the middle of the visible spectrum and so seemed brighter.

But if you're not taking advantage of that effect, then I don't
think pulsing gets you anything. The instantaneous output of the LED is
proportional to the instantaneous current, and the percieved brightness
of the pulsed light source is the same as the average brightness, if it's
pulsing fast enough ( 10 Hz or so ). So it works out that the percieved
brightness is the same as it would be if you'd driven the LEDs with
a constant current.

If resistive losses are significant, then pulsing will actually reduce the
overall efficiency, since resistive losses are proportional to the
average of I^2.

On the other hand, if your battery is at a higher voltage than you want
to drive the LEDs, then pulsing them might be more efficient than using
a linear regulator --- that has more to do with the inefficiency of the
linear regulator than anything to do with the LEDs, though :-)


I did some googling to refresh my memory about pulsed LEDs and found
a short article on the subject:
http://www.caves.org.uk/led/foot4.pdf

On the other hand, I also found some articles on pulsed LEDs saying
the opposite --- that the percieved brightness of a pulsed source
is closer to the peak brightness than to the average brightness.
It might be worthwhile tracking down some actual experimental data,
or doing an experiment yourself...

--
Wim Lewis , Seattle, WA, USA. PGP keyID 27F772C1

If the brightness of an LED is a 1:1 linear relationship with the power
supplied, that suggests that perceived brightness of a pulsed LED would be
neither at the arithmetic mean (average) nor at the peak but rather at rms
(root mean square). Comments?