??? 07/18/06 12:27 Read: times |
#120459 - not all of us Responding to: ???'s previous message |
If you want to know why some people think the 8255 a poor choice, I'm sure there are lots of opinions available here, though I'd also guess that most are from people who have no real experience with these very flexible and useful devices
I have lots of experience with the 8255 and find it a total anachronism. Yes, in those distant days when the uC itself was not much more than a CPU and speed demons could handle 8MHz it was 'a solution' that could be wrenched in. Yes, it "worked" and did the job, but even when 12MHz at 12 clocks became available it started falling by the wayside. Today you have uCs with tons of internal memory etc, so why not make use of the multitude of pins you get when no external mamory is used (not to mention 8 porters) instead of wrenching external logic onto the board. My progression has been from 8255 to CPLD to nothing with the occsional external HC chip. Today I achieve with a SILabs 8 porter what once took 4 or more chips. If what this does provide is not enough there is the 400kHz (now 1MHz) IIC bus and the even faster SPI bus, not to mention the simple chain of shift registers and, of course, the CPLD/FPGA. Even the lowly sub $1 LPC chips have more 'I/O devices' inside than a 8255 can provide. If you need 'a lot' look at the offerings from e.g SILabs and Ramtron. Erik a note to Richard: the Ramtron devices offer the clocking schemes for the UART you have been looking for, |