The thing that worries me about the server/client approach (although master/slave or primary/secondary terminology is more accurate) is that the amount of processor horsepower is limited, which in turn limits both the complexity of the display pattern and the number of downstream nodes on the network. I'm thinking that the firmware complexity increases if the outgoing packet takes longer than 1/120 sec to transmit (for 60Hz AC line frequency).
P. Short wrote:@budude - while it isn't there now, I intend to incorporate the ren4flood changes into this code. It's only a few lines of code to do the r4f changes.
Awesomeness -- many thanks Phil!
P. Short wrote:@erm213 - Not a lot - the eeprom is only 256 bytes, the data RAM is similarly limited, and the program memory is not writable. The data patterns would have to be algorithmically generated with a small amount of custom code, it's not really practical IMO to download part of a vixen sequence into the part (besides, there isn't an interpreter available for Renard now).@tjetzer - that's possible, if someone is up to writing the code for that. Are you more interested in having each PIC run independently, or have server/client type of setup? Bear in mind that the clocks are not perfectly synchronized, so even if the patterns will end up out of phase, even if they started out in sync.
Not a lot of room, except for a very small sequence. Oh well. Thanks for the reply.
Erik
I suspect that Mac or bcstuff will request some change, and I'll add the r4f change at that time.
Just a thought on this one. It would be good to have either a standalone pic with serial in and out, or and option in the base renard firmware that when no serial is received for a period (say of over 1 minute) that all outputs are turned on, so that if the PC crashes or is not available for other reasons I can still see my lights.
Rob
