Don V's BYW (BTW?) gives one of the reasons I put 1N4148 or 1N4150 signal diodes in series with the LEDs on my RR. Mine are multiplexed, so spend 7/8 or their time actively driven in the reverse direction.?

The diodes block the reverse voltage and current, reducing the loads on the drivers and any dimly glowing LEDs that are supposed to be off, glowing because of current leaking through reverse biased LEDs. Some people question whether this is possible, but I had the problem and the diodes fixed it.

When LEDs are scattered around the layout, whether on panels or in signal heads, or anywhere, I learned the hard way that they sometimes fail from voltage transients. To prevent that, I also put a reverse biased signal diode in parallel with each LED. Problem solved. I'm not aware of anyone else reporting that problem. YMMV.

Don W

--
Don Weigt
Connecticut