Page 1 of 1

LED matrix display from SK6812RGBW LED strip

PostPosted: Mon Jun 25, 2018 7:54 am
by ric
Just a "proof of concept" test to see if this was possible...
5 metre roll (60 LEDs per metre) cut into eight strips of 37 LEDs.
Serpentine connection to keep wires short, corrected in software.
Edit: Controlled by a PIC16F1829 running from the internal 32MHz oscillator.


Re: LED matrix display from SK6812RGBW LED strip

PostPosted: Tue Jun 26, 2018 3:41 am
by AussieSusan
And supporting our local public broadcaster as well!!!
What are you driving it with?
Susan

Re: LED matrix display from SK6812RGBW LED strip

PostPosted: Tue Jun 26, 2018 4:18 am
by ric
Oops, I mentioned that in the Youtube comments, but not here.
It's a PIC16F1829, but with custom labelling so you can't read it off the chip. ;)
I'm using the SPI peripheral for timing, running the PIC at 32 MHz, SPI at 32/4 = 8 MHz. One byte per transmitted bit.
PIC16F SPI won't run faster than one byte every 12 clock cycles, so that's 8/12 = 666 kHz, or 1500ns per bit so not quite the optimal 1200ns per bit.
The SDO pin is directly driving the Din pin of the first LED.

I considered using the USART in synchronous mode, but I need it to make this into a general purpose display with serial receive.

I see DavidBLit did a similar thing in 2014, but achieved a totally different way on another PIC16F1xxx device.

Re: LED matrix display from SK6812RGBW LED strip

PostPosted: Tue Jun 26, 2018 7:50 am
by ric
AussieSusan wrote:And supporting our local public broadcaster as well!!!
...

Total accident of which frame Youtube grabbed as a thumbnail.
I could add a crude Lissajous graphic though ;)

Re: LED matrix display from SK6812RGBW LED strip

PostPosted: Wed Jun 27, 2018 2:16 am
by DavidBLit
Very nice work! I've been thinking about doing something new with the SPI-driven strips but haven't yet jumped in. The most-recent project used a PIC16F1825 from the 2014 stuff to generate the WS2812 bitstream while accepting color commands via UART (from an Apple II+, of all things ;-) It only took about a second per LED for the RGB calculations on the Apple II+. :lol: