Hi,
Snunn - glad you got things working. Here is an interesting post on language / speed benchmarking test - but doesn't include java (tbh - all a bit old now). http://codeandlife.com/2012/07/03/benchmarking-raspberry-pi-gpio-speed/
I'm heading down a slightly different route, as my 74HC595's have turned up tonight.
Had a nice little play - for the 3 wire pixels (if you use all the same type) these should be cool.
The 4 wire are a little more difficult - as I only want to use spi - do you send data twice (bit one being data sent twice and bit two being clock sent once high and once low - advantge, doesnt mess with the data, but halves the number of pixels) or do you use hardware, and allways send 8 bits then use a binary counter (changing on 4 cycles) to provide the clock for the pixels (chances of error with missed packets).
Anyway time for bed - thats one for next time.
I have been able to "route" the clock and the data line of the SPI output to a desired output string. I am using 2 - 74HC4514's and a single 7404 invertor chip. This allows
me to multiplex a single spi output(clock, and mosi) to one of 16 destinations. The last thing is to use the GPIO pins for the Latch Enable(1 pin) and address selection (4 pins) and connect those to the 4514's. My question now is, I am not a hardware guy and don't know what all resistors I should have in play to keep the signal noise down...
any ideas?
I'd probalby suggest you use a tri state buffer on the outputs.. Something like a 74HC540 or 541 ( depedning if you need inversion or not ).. And stick a 100ohm resistor in series with the output. This
For those that are not followign this, this project could be the killer proejct for 2013 in the DIYC space.


