I'm facing a conceptual dilemma, many of the gurus here and elsewhere =
recommend to keep interrupt code to a minimum. That's what I'm trying to =
do, but I'm having some serious problems to transfer data gathered at =
the interrupt procedure and use it on the main loop.

The interrup is receiving a constant stream of UART data at 19200baud =
(it will be faster as soon as I get the appropriate crystal). What I =
have to do is to interface with my compactflash card and store each one =
of those bytes in it. I intend to store data as FAT files (I know, 512 =
bytes blocks). Thing is when I'm trying to process something on the main =
loop, things are coming at a faster rate on the interrupt.

One thing I've tried was to setup a circular list of 64 bytes, push data =
on the interrupt and pop it on the main loop, it didn't work. Then I've =
increased the circular buffer size to 128, then I started to see some of =
the expected bytes on my CF (actually I'm sending them via USART to the =
PC just to test, but I'll replace to writing to the CF later). When I =
increased to 255, then almost got it, but still the interrupt was =
winning the race. I feel that if I had a big enough buffer, eventually =
it would work out, but perhaps there is a better way to solve it.

Should I put all my CF writing code in the interrupt procedure? Would I =
be risking to get overrun errors?

I have to confess, I'm enjoying learning all these new things about PIC =
and electronics, but sometimes it's frustrating, things don't happen at =
the same speed of software development... or I'm too impatient to be a =
newbie again.

Thanks again!

Padu
_______________________________________________
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist