Re: [ATM] LED's and Diode Electronics

["matt" <electro_optic@bellsouth.net>]

1- forget the solid state physics for your practical application .
 If you just want to further your theoretical understanding and see more
details about diode conduction, there are reasonable web references , for
example:
http://ece-www.colorado.edu/~bart/book/book/chapter4/ch4_4.htm
which treat it in some detail and explain the various conduction mechanisms
, together with their equations, solving and approximate solutions or curve
fitting .

2- do not attempt to feed LEDs from large caps unless you actually intended
to create SSEDs (spark and smoke emitting diodes ).
Exceeding certain local current/carrier densities through the diode junction
leads to irreversible destructive effects, even if it's a short duration
peak .

3- now the practical part - energy efficient LED power supplies need to
regulate the _current_ to the desired optimum current for that particular
LED type. Not a voltage regulator but a current regulator .
See LED manufacturer data sheets to determine which current value you're
aiming for, then all you need is a simple switching _current_ regulator .
For the past decade, such circuits have been available on a single chip,
requiring only a couple of external components (typically 2 caps and 1 small
inductor)  . You may buy all these parts including the switching regulator
IC for just a few dollars (shipping would run more than  the parts total )
from distributors such as Digi-Key .

best regards,
matt tudor


-----Original Message-----
From: Richard Schwartz <richas@earthlink.net>
To: atm@atmlist.net <atm@atmlist.net>
Date: Saturday, June 18, 2005 12:02 PM
Subject: [ATM] LED's and Diode Electronics


Last night's PPC/CHHU meeting was mostly about LED's.   Richard Nelson
brought some samples that he picked up at ALL Electronics, and we looked at
the I-V curve of one such device.    The equation for current through a
diode is I = Io*(exp(mE/nkt)-1), but I am not certain if that is the
appropriate equation for a LED.    LED's seen to operate at higher voltages
than ordinary forward-biased silicon junctions.   I guess the first thing to
do is curve fit the data we have...

Particularly, there was interest in bright LED flashlights, and the question
came up of how to supply power to a LED.   The usual method is through a
current limiting resistor, and that is wonderful for you if your name is
Eveready or Duracell.   But why not a capacitor?   The initial current would
be very high, but it would be for only an instant.   The idea is to put all
of the power into the LED and not put some of it into a resistor.

Along with this question, there is another one:   How does the photon output
of the LED vary with current?    You need to know this to calculate the flux
during a capacitor discharge through the LED.  You want to choose the
voltage, capacitance, and firing frequency  to deliver the most photons per
joule of energy while not overstressing the diode.   All of this is highly
dependent on temperature.

Finally, note that white LED's use blue light from the semiconductor to pump
a phosphor that emits white light.   There is a noticable time delay in the
phosphor that renders a white LED unsuitable for applications that require
high frequency modulation.

Hopefully somebody will get interested in this and we will have brighter,
more battery-efficient testers.

. . . Richard
_______________________________________________
ATM mailing list http://www.atmlist.net/


_______________________________________________
ATM mailing list http://www.atmlist.net/