I have a simple LED circuit (2 strings of 2 LEDs each with a ballast R). Input to the circuit is from LED driver (60mA). There is no protection diode at the input/output due to load voltage requirements. LED reverse protection is provided through voltage divider as max reverse voltage is -12V. The LED has a max reverse voltage of -6V. If a cap is put from driver input to ground, will the circuit survive a -300V pulse lasting 0.05ms or is a reverse protection diode needed?

Thanks for the help.

 

If the cap is between the driver output and the limiting resistor, maybe. LED's don't like reverse bias, so a protection diode is a good investment protector.

 

A diode was present in the original design without the resistor network, however the voltage drop of the diode put the load voltage of the circuit over the max. The load voltage is used as a feedback voltage for fault indication. Another reason the diode was removed is so a small resistance (ballast resistor) could be used to balance the current division in each string instead of the forward voltage of the LEDs. Some of the voltage gained from removing the diode is now developed across the ballast R. The circuit needs to be immune to transient waveforms like the one described above. I have attached a sketch to help illustrate.

Input is always appreciated.

 

Resistors in parallel with the LED's? That looks awkward to arrange. Do you need to exactly balance irradiance? I would think a couple of LM317's set up as current regulators would be easier to get good results.

 

I've seen a similar scheme, with 4 or more diodes for a really simple bargraph (math is slightly complex though). Doubt that is what the OP is going for though.

 

The balance of current is acceptable in the present config. The four LEDs are used to illuminate an icon and need to be uniform in appearance. There would be to high of a cost increase to add the LM317(s).

The main issue is that LEDs do not like reverse voltage(s) that exceed the max rating even for a short period of time.

A real easy way of confirming EMC immunity would be to test the present config however, I do not have the wallet to be wrong so i am trying to get some advice/knowledge to help with the decision. I am quite certain that if there was a diode on the input line with the presence of C1, it would pass EMC as any waveforms would shunt to ground through C1 and the diode would not allow a large (-) voltage to develop. I am unsure of maintaining EMC immunity with only C1.