| Electronics > Projects, Designs, and Technical Stuff |
| kCube - LED cube design revisited |
| << < (5/8) > >> |
| tatus1969:
--- Quote from: cowana on March 28, 2018, 08:58:44 am ---It's hard to see in your 3D model of the flexi PCB design - are the 100nF decoupling capacitors mounted on the underside beneath each LED? --- End quote --- The flex PCB is single sided to reduce cost, and there is only one 1uF capacitor at the end of each leg. The WS2812B chips have an internal constant current source delivering roughly 30-50mA, which is then PWMed. The second LED on the leg is the most critical. As there are 8 on each leg, the second supply trace (0.2mm wide, 13.6mm long) will see a voltage ripple of ( 7*50e-3 ) * ( 0.0171 * 13.6e-3 / 0.2 / 0.035 ) = 0.01V. The trace inductance will add more switching noise, but my first tests didn't show any problems. |
| tatus1969:
Now I got worried over my own explanations, and did another experiment. I removed *all* capacitors from this 24-LED ring. And it still worked flawless (!) I measured 1.5V switching spikes at the last LED, which coincides with 0.3*VDD and suggests that the system is probably just about to fail. Adding one single 1uF MLCC at the first LED reduced the spikes to ~ 250mV. |
| tatus1969:
The first parts arrived, and I am curious of how that will work together 8) The flex PCBs are also being manufactured, and I have got an interesting information. The base material will be PET, which is great because it surely is transparent, and its refraction index (1.585) almost matches that of acrylic glass (1.49). The PCB tracks will be immersion silver coated. |
| tatus1969:
Just got a picture from the PCB fab, this is a reference picture that shows how my PCBs will look like: |
| ataradov:
Sol there is no cutouts, it is just a continuous sheet of plastic? That will also add up in opacity over 8 layers. |
| Navigation |
| Message Index |
| Next page |
| Previous page |