Need reverse diode on SMD LEDs?

[mmagliaro]

I am considering using a tiny surface mount LED in a headlight for my loco project (like an 0402).
For the folks who use these regularly, do you wire a reverse diode across the LED, so that spikes from
the motor during brief moments of intermittent contact don't blow up the LED?

This was a problem with white LEDs for a long time, even the larger "normal" sized ones like T1's.   The problem is
that during intermittent contact, the motor is shut off for an instant.  As the motor field collapses, it produces a
pretty high-energy spike, and white LEDs would sometimes burn out from it.   You can see this if your loco
has a back-up light.  As it's moving forward, if there is a bit of dirty track, sometimes you will see the back-up light
flicker on for a moment.

Anyway... do the tiny surface-mount LEDs blow up from this?   

Thanks.

[SkipGear]

I haven't blown one up because of the voltage spike but I put a diode inline anyhow just to keep the lights from flashing durring dirty track. A little bit of precaution is worth it at this point. I hate doing work twice.

[kornellred]

Perhaps powering the SMD LED using a bridge rectifier would serve the same purpose in protecting it from back EMF and save a lot of soldering hassles in the process.  It could possibly even eliminate flicker in certain situations, and the LED will illuminate in reverse!

[jcox3751]

From my friend at Richmond Controls:
 
"My OPINION is that every WHITE LED needs some form of protection between the LED and the motor, IN ADDITION TO A CURRENT LIMITING DEVICE LIKE A RESISTOR.  Something like a Richmond Controls circuit or a decoder takes care of the problem.  With no such circuit in place, I recommend using a FAST diode and a small ceramic capacitor.  A SLOW diode, like a 1N400x does little or no good.
 
This topic is discussed at www.richmondcontrols.com, in the LEDs section, in a document near the end titled "Recommendations for Using White LEDs".
 
I can't think of any way that a reverse diode could help protect against flickering.
 
I've never seen a case where a colored LED (red, amber, green for example) had this issue, although I'd be careful with green and blue LEDs.  The colored LEDs appear to be more robust."

From Jim Hinds

[mmagliaro]

Yep.  I have read that before, and I have experience white LED burn-out from the reverse EMF spike from the motor.
The reason they specify a Schottky barrier diode instead of a conventional one is that the reverse diode has
to have a switching time and forward voltage that is lower than the LED, or it cannot possible switch on fast enough to short
out the spike before the spike kills the LED.    (For anyone else reading along....  we usually don't think about, or worry about,
"switching time" with diodes.  We tend to think of them as rectifiers that just conduct in one direction and block in the other.
But at the instant a voltage is applied across a diode, there is a brief instant where the diode has to turn on before
it starts to conduct.  In this case, that turn-on time is critical.)

What I wasn't sure about was whether SMD LEDs had this same problem.   It sounds like if it is a WHITE LED, it would.  And
it's not worth the worry, so I will put in the barrier diode.

[peteski]

SMD LED is electrically no different than any other LED. The actual semiconductor chip is the same as in LEDs in any other package. The only difference is the size of the package itself.

White, blue and certain other newer LEDs have in fact quite a low (destructive) reverse voltage.  Much lower than the older red, green, or yellow LEDs.  Those voltages (as low as 5V) are all stated in the technical specs for each LED. But again, the package size is not really related to  the electrical properties of the LED itself.

BTW, shouldn't this thread be in the forum dedicated to DCC and electronic related topics?