Hello All,

I have Five LEDs in series (3.2V @ 20mils typical) in a circuit as shown in the attached file. The input is ~24VDC, with 680 ohm resistance in series with the LEDs. LED Specs: http://superbrightleds.com/specs/w110-360_specs.htm

What could cause the 2nd and 4th LED's to short? My first thought was that the LEDs were bad. I swapped in a new board. Within a few weeks, the 2nd, 3rd and 4th LEDs shorted. I am using these LED boards on seven other machines, none of which have had this issue. Could the environment effect LEDs?

Something to note: the LED board can be manually turned on, for testing. In this case, the machine it is connected to is actually off. When the machine is running, the input to the LED board drops upwards to 0.5V. Perhaps the LEDs the shorted are more sensitive to this voltage drop than the other LEDs?

Any insight will be appreciated.

 

Well, it seems you've rated them conservatively enough.
As it is, the 680 Ohm resistor is only allowing about 11.8mA current flow when +V = 24v.
(24v-(5 x 3.2v))/680 = (24-16)/680 = 8/680 = 11.8mA.

You may be having problems with transients, particularly if the machines have large electric motors. Connecting a diode in reverse across all of the LEDs should help that.
[eta]
Here's a couple of examples:

On the left, D1 clamps reverse EMF. Turn-on time for the 1N400x series is fairly low.
On the right, D2 protects against reverse EMF up to 1KV. However, the relatively slow switching times (turn-off) of the 1N400x series may not be fast enough to prevent damage to the LEDs.

 

Well, it seems you've rated them conservatively enough.
As it is, the 680 Ohm resistor is only allowing about 11.8mA current flow when +V = 24v.
(24v-(5 x 3.2v))/680 = (24-16)/680 = 8/680 = 11.8mA.

You may be having problems with transients, particularly if the machines have large electric motors. Connecting a diode in reverse across all of the LEDs should help that.
[eta]
Here's a couple of examples:

On the left, D1 clamps reverse EMF. Turn-on time for the 1N400x series is fairly low.
On the right, D2 protects against reverse EMF up to 1KV. However, the relatively slow switching times (turn-off) of the 1N400x series may not be fast enough to prevent damage to the LEDs.

Thank you for the response. I'll have to get some diodes to test this. I also got an oscilloscope in today, I will monitor the transient for any surge's.

 

Do you think adding a capacitor after the resistor, in parallel with the LEDs will work?

 

Adding a cap may very well help.