White LEDs tend to require a higher average current and have a higher
voltage drop (for equal average klight output) when pulsed than when
operated continuously. Unless the average current is less than a few
milliamps, you get more light and less heat with continuous operation than
with pulsed operation at a frequency high enough to apear continuous.
By and large, only LEDs that are more efficient at higher instantaneous
current benefit from pulsing. This was especialy true of LED digital
displays with GaAsP on GaAs substrate, where a segment had an average
current of only a few milliamps but the efficiency was maximized at
instantaneous currents of 50 milliamps or more, often at least 100 mA.
Many people were not aware of this nonlinearity of those LEDs and believed
that the benefit of pulsing was due to a quirk of human vision.
Some LEDs, namely at least some InGaAlP ones, have a degradation mode
that is a function of temperature and duty cycle. I suspect this is from
some sort of diffusion of an ingredient from where it belongs to someplace
else and dependent on electric field around the boundary or zone that the
diffusion occurs across. Maybe a boundary between different layers
(having/lacking whatever diffuses) should be sharp but gets "blurred".
LEDs with that chemistry, primarily at higher temperatures and lower
average currents, can have a life extension from pulsing. One
manufacturer, Agilent, publishes an application brief where they encourage
pulsing of their LEDs with that chemistry if the average current is less
than 10 mA, and recommends instantaneous currents 10-100 mA (but average
current not exceeding 30 mA).
I discuss pulsing LEDs in an attempt to make them brighter (or appear
brighter) in a web page of mine,
http://www.misty.com/~don/ledp.html
By and large, if the average current is already close to the maximum
rated constinuous current, there is at beast little to gain from pulsing.
- Don Klipstein (
[email protected])