These are common cathode and the forward voltages at 20mA are:
Red 2.2v
Green 3.3v
Blue 3.3v
From this, I assume that I don't need a resistor on any leg but the Red when powered from a 3.3v system. And on the Red leg, calculated resistance is 55ohms. Is this a safe assumption.
Part two:
I need to run a potential total of 7 of these for my project and would prefer to stay with the 328p. The obvious problem here is that I would need 21 pins and a lot more current than the 328p can support. So what is the best solution? I do not need PWM so simple is better. I also will never use color blending. They will either be red, green, or blue so I would never have more than 7 LED's powered at once plus one for my illuminated switch.
Any suggestions on a tried and true method for this? I know there are LED drivers but I don't need PWM so my feeling is that an LED driver is going to be somewhat redundant. I considered a parallel out shift register like the 74HC595. Everything about this seems to be tailor made for my potential 8 LED project except of course that it would be limited because I need a total of 24 outs. So three of these???
I found options for more pins on outputs on digikey but everything appears to be sinking which obviously won't work with common cathode LED's.
From this, I assume that I don't need a resistor on any leg but the Red when powered from a 3.3v system. And on the Red leg, calculated resistance is 55ohms. Is this a safe assumption.
With +3.3v you will probably have problems with the Green and Blue (3.3v) LED's because of their forward voltage drop. +5v sounds like a better choice. In any case you still need one limiting resistor per anode suited to their respective fwd voltages.
Everything about this seems to be tailor made for my potential 8 LED project except of course that it would be limited because I need a total of 24 outs. So three of these???
Yes three chained together would be fine.
Do not put a capacitor on the latch pin, that is an error. Put in instead between +5V and ground close to each chip.
The forward voltage of LEDs depends on the temperature, the current flowing, the manufacturing process and so forth. Expect the actual forward voltage to be +/- 15% of the nominal voltage from the datasheet...
And drive LEDs with (limited) current, not voltage because the temperature dependence can lead to thermal runaway otherwise (the forward voltage for a given current falls as the chip heats up (so the current for constant voltage increases).
The figure you really want from the datasheet is the minimum forward voltage for the temperature range you are interested in - you might find a graph of forward voltage against temp if lucky, but its probably average not worst-case...
The other approach is S.o.T. (select on test) - try different resistors to set the current for the actual device as you assemble the circuit...