On Sat, 11 Nov 2006 06:59:18 +0000 (UTC), (Don
Klipstein) wrote:

In , Mark Lloyd wrote in part:
sylvan butler wrote:

Of course, if you were making a string of LEDs intended to run from
A.C., you might also wish to use a series diode in addition to the LEDs
themselves. Why? Because a normal rectifier diode has a much greater
tolerance for inverse or reverse voltage than does a normal LED. If the
peak-inverse-voltage (PIV) specification of a diode is exceeded, the
diode will usually be destroyed.

I once thought so. It's not true. BTW, if it was, it would make it
risky to test LEDs of unknown polarity.

Reverse biasing will not hurt a diode (LED) as long as the proper
series resistor is present. It limits current to a safe value.

Note that exceeding the PIV causes "breakdown" which just means the
diode starts conducting. It is not damaged unless the current becomes
excessive. Zener diodes are operated this way (in breakdown) all the
time. Right now, I have a device (part of a holiday light controller)
that has 2 LEDs on RS232 outputs. these switch polarities (+/-), so
the LEDs are reverse-biased half the time. They haven't been harmed
yet (during about 60 hours of operation).

Some LEDs are intolerant of reverse breakdown. GaN and InGaN ones (most
other than infrared, red, orange, yellow and yellow-green) can be
significantly degraded, sometimes to the point of producing zero light at
a significant fraction of normal current or even zero light at normal
current, by reverse breakdown. It appears to me that some localized
electrolysis process or microscopic localized heating causes a partial
short. This damage occurs even from static electricity, and many of these
LEDs are considered static-sensitive.
Thankfully, in my experience at least, these LEDs tend not to get
damaged by exceeding their typically 5V reverse voltage limit by small or
moderate amounts. I can usually get away with testing one with a 9V
battery (and appropriate resistor) with wrong polarity because in my
experience they usually need more voltage than that to break down.

I suspect that at least some other LEDs can suffer localized heating
effects within their chips during reverse breakdown, and in that case
would not tolerate full rated power input during reverse breakdown. At
least with LEDs other than ones with GaN or InGaN chemistry, I don't hear
of this being so bad as to have an unnoticeable zap of static electricity
causing actual damage.

- Don Klipstein )

So if I wanted to test a LED with a 9v battery, how do I know which
leads on the LED to connect to the POS and to the NEG on the battery,
and what size resistor do I use? Does it matter if the resistor is on
the POS or NEG side?

Also, reading what you said, does this mean that if I connect the LED
backwards on the battery it will burn out?

Thanks