Has anyone discovered any adverse results by adding capacitors to APA102s? I know the datasheet does not mention them, and anything that comes out of China won’t have them either. However I’m skeptical still and would rather be safe than sorry.
Why exactly do the WS2812’s require a capacitor?
Voltage ripples.
I think that beefy power caps are always useful, but for whatever reason they have no need for the small decoupling caps you usually see on WS2812 strips. I think the controller chips in each LED are more tolerant of minor rippling in the signal lines.
But if you want to add decoupling caps, I don’t think it can hurt.
@Ashley_M_Kirchner_No I have not yet played with the APA 102 but I can assure you that you would never get any adverse effect by adding small decoupling caps on the DC voltage supply wires of your APA102s.
I also question the absence of caps on the Chinese APA102 strips. I believe that is driven purely by cost considerations with disregard of sound digital circuit design principles.
Yeah I didn’t think so either, but not having any to play with at the moment, and designing a circuit for them, I am planning to add them in. Just wanted to hear from folks who have had a chance to play with them and if they added caps and noticed anything go wrong.
@Ashley_M_Kirchner_No If I was designing a circuit out of the APA102, I would definitely add them, no questions about it !
Please believe me, you will never have an adverse effect.
One cap per device would be ideal! Note that you could also distribute the caps, 1 cap per 2, 3 or more LEDs.
I think the only possible downside to adding bypass caps would be the combined capacitance on the power supply, making for “large” inrush currents. But even if you had 100 0.1µF caps, that’s only 10uF of additional capacitance. If your power supply can’t charge that you’ve got bigger concerns. 
The only reason I bring this up is I’m working on some bus-powered USB designs, and the USB spec says you can’t have more than 10uF to limit inrush current.
@Tod_Kurt You’ve probably already seen something like this, but I remember a simple circuit in an FTDI app note for limiting inrush current. You could control a FET based on the state of USB enumeration. http://www.ftdichip.com/Support/Documents/AppNotes/AN_146_USB_Hardware_Design_Guidelines_for_FTDI_ICs.pdf
section 2.4.2
@Tod_Kurt That is a valid point specially in the case of USB that is likely powered when you connect to it and USB power does have low practical current limits.
In the case of a normal PSU, they often have 100uF or much more already on the output so a few hundred or even thousands 0.1uF caps is not significant.
@Pixelmatix I’m using that technique, but wow I wish I would’ve seen that doc before doing a bunch of research on it. It very nicely explains the problem and how to solve it. Bookmarked. Thanks!