Before I dive into this, did anyone try to implement parallel output for SPI chipsets using FastLED?
I want to frankenstein the ws2811 parallel output with SPI bitbanging to get multiple data lines sharing a single clock line (multiplexed in a breakout board).
Thanks!
Which board? There’s instructions on github for the teensy 3.2. The same technique might work on other boards
I did a quick try but the aggregate data rate was stupidly low and tossed the code - especially once I figured out a way to be able to drive four apa102 lines at hardware SPI speeds on the teensy 3.2 (they weren’t in parallel but because each line could run at 24mhz I was still getting far higher throughput than any parallel attempts got).
This is for the esp8266. I want to write to the 4 pins on PORTA and strobe the clock between writes. Unless I’m missing something - shouldn’t it be about as fast (or slow) as a single bitbanged SPI output?
In this case, the bulk of the CPU cost is going to be the scaling/dithering of pixels - not the pin strobing. If you’re driving 4 apa102 strips calling addLeds four times - each with a different data pin and all with the same clock pin would take about the same time to write out a frame as any hacked up “parallel” output would take. (It might be a bit slower, but not sure that it’d be slower enough to justify the code juggle to make the output parallel). (Because apa102 strips open their frames with 1’s if you toggle the clock and the data line is held to 0 then that strip won’t see any updates - which makes it easy to share clock lines with the apa102 for multiple output)
Awesome, thanks! Thats enough for me to complete the board layout