I have further tested the HC4067 16 Channel Multiplexer to drive multiple APA102 strips.

Archive FastLED Archive
,
1

I have further tested the HC4067 16 Channel Multiplexer to drive multiple APA102 strips. I have successfully driven 2 strips of 12 pixels at 20Mhz using an Rpi A+. This $4 chip can help add redundancy to large projects, and lets you drive APA102 at higher speeds when you have more then 500 pixels(on strips). You can chop up a huge layout of, for example 2000 pixels into blocks of 200, and drive them at speeds higher than you would be able to drive 2000 pixels in a long chain. Example code here https://github.com/leonyuhanov/rpi_apa102driver and little video here of it working https://www.instagram.com/p/BEh_8g4RDVP/

6f0aaffddfa005bb51f32e0c92ae5a3c.png
6f0aaffddfa005bb51f32e0c92ae5a3c.png853×737 154 KB

2

Just so you know, you’ll get a longer maximum transmission length if you use a 74HCT245 buffer chip instead of using a ghost pixel (a single APA102 repeater will get you about 1.2m further along Cat5, while a 74HCT245 will get you 3m).

Also, keep in mind that if you have a whole heap of data lines in parallel with one output, the total length of all of those lines will affect your maximum transmission speed (you will come to curse capacitance on data lines). This isn’t a problem if your controller is right next to your LEDs, but if you’re trying to push data through 10m of Cat5, you will quickly learn that raw TTL signals aren’t enough.

3

Thanks for the tips

4

@Luminous_Elements Have you personaly used the 74HCT245 with SPI running at above 10Mhz? Do you know what that ICs maximum frequency is?

5

I ended up using 6MHz through the 74HCT245 (on one of @PaulStoffregen 's OctoWS2811 boards) for the Tripper Trapper. I had to lower the speed because of the huge capacitance on its long lines, but I have run signals through the chip at 12MHz through about 50cm of wire to a 60px@30LED/m strip (which was the test rig for that project).

Datasheet says transition rate is 1.67ns/V [1]. At 5V this is 8.35ns. At 12MHz this means the low-to-high and high-to-low transition time is around 0.1 clocks. I’m not sure what to make of that, but the APA102s in the test rig seemed to read it well with the short lines. Again, at higher output capacitance, this goes out the window.

[1] http://www.nxp.com/documents/data_sheet/74HC_HCT245.pdf

6

@Luminous_Elements Thanks a million for that info mate, its hard to find actual results!

7

Don’t I know it :expressionless: