What Daddle said. The problem is that there are three ways to provide the constant current
source needed by LEDs, none of which are ideal.
The old-fashioned method is a series resistor, but this is only useful when there is a constant voltage source (not a boat with battery-charging fluctuations). Also, it's extremely wasteful.
The second, which it sounds like you have, is a switching power supply (or, sometimes, just pulse-width modulating the LED). Either of those involves lots of sharp on/off transients, which create harmonics into the hundreds of Megahertz. (Technically, a square wave is a monotonically decreasing sum of all odd harmonics... I had one in a homebrew LED bicolor circa 2000 that completely blanked marine VHF
in the vicinity.) Without some good engineering, this is basically a big sloppy transmitter, and it's not easy to go back after the fact and provide enough filtering to fix it. The advantages are high efficiency and small/cheap parts
The third method is also a constant-current source, but it is done linearly. This costs more power, but does not generate noise
. An operational amplifier or other active device controls a "pass element" (a field-effect or regular transistor) to adjust the current
provided to the LED, keeping it at the same level regardless of applied voltage.
For the manufacturer of yours to suggest a tiny cap like .1 and then suggest you "try other sizes" is really cheesy!
The trade-off that we get down to is either a well-designed switcher (best power performance) or accepting the lower efficiency of a linear regulator
. Fun. eh?