I'm using a 5V buck regulator to power an LED (yes, I know I should be using a constant current LED driver).
I have resistors inline with the LEDs maintain 150mA per channel (RGBW).
The 5V buck regulator spikes to 7.5V for a fraction of a second when turning it on and it ends up burning out one channel of the RGBW LED.
I think I should be adding inductors inline with each channel to prevent that sharp change in current from burning out the the LED.

How do you size an inductor to prevent that brief increase in current?

 

Inductors are bulky and expensive, and you'd have to measure the duration of the spike to know how to size the inductor.

Is the buck regulator built or purchased?
Do you have a circuit diagram for it?
The best why would be to determine why the regulator is spiking and/or reduce the turn-on turn to eliminate the spike.

 

Inductors are bulky and expensive, and you'd have to measure the duration of the spike to know how to size the inductor.

Is the buck regulator built or purchased?
Do you have a circuit diagram for it?
The best why would be to determine why the regulator is spiking and/or reduce the turn-on turn to eliminate the spike.

It is a pre-built Remcom Power's R-78E5.0-1.0.
I suppose I could put a single inductor inline with the output of the buck regulator.
Still not sure how to determine the size of the inductor though.
I might just buy a variety and kind of guess at it.

 

Use an Inductor, and a Capacitor,
based on the formulas in the Data-Sheet of the Switching-Regulator that You are using.
Like this .............
.
.
.

 

Yes, and put that zener diode at the LED side of the spike suppression filter. And also have, in addition, some 0.1 mfd capacitors across the LEDs, because they have less series inductance than the big electrolytic capacitors..

 

Leave it alone, and make a simple constant current driver for the LEDs. Then a brief over-voltage won’t do it any harm. You are already driving the LED with a transistor, you might not actually need any more components to drive it constant current, and you also lose the problem of thermal runaway when the forward voltage drops as it get warmer (resulting in higher current, more heat, further voltage drop etc.)

 

Below is the LTspice simulation of an example, simple two-transistor constant-current driver, as Ian0 suggested:

Note that the LED current (green trace) is quite constant, from the initial 7.5V transient (red trace), then down to the 5V normal voltage, so the transient will have no adverse affect on the LEDs.