One would parallel-wire strips of 32 or 64, thus requiring 16< parallel data and latch outputs..

stabilizing some of the lines, active repeating.. would also be an addition to stability

I’m very disappointed with behavior of those chips when a little of radio electrical noise or other EMC trash enter to it, they go completely mad, especially if linked in chain….
then nothing other help then re power… :-(
for those project, this is trashed, money….
pcb, was made pro. with good power and signal traces, with many capacitance for stability, so this is not design problem, but very bad designed chip…… (with poor support of allegro) ok it work for hobyst, but for larger pro. project fail.

I believe it was also his site that introduced me to Sparkfun. (As well as his eBay supplier for the RGB LEDs on the shiftbrites)

I see he now has a Shiftbrite shield for the Arduino, I’ll have to add that to my to-buy list. :)

the problem seems to be that the output of each individual red, green and blue LED does not go low enough with PWM to produce the very low levels of each colour required for more subtle shades.

the 10 bit PWM on the allegro is pointless, you will be lucky to get 4 bits of colour resolution from an RGB LED. it may seem like more in the demos where they fade from one colour to another, but that’s only dealing with primary colours which are some combination of full R/G/B.

the problem is that even at the lowest end of PWM, the current required to turn the LED on is too high to get down to under about 40% brightness with a 1 in 1024 (10 bit) PWM. Any lower and the LED does not light at all.

the only thing which may work is constant current control, but i have not had time to experiment with it yet.

Shame there’s no red ones, especially at that price!