Erik,

It's a Universal Asynchronous Receiver/Transmitter.

Since the 805x serial channel is not only Asynchronous, it's not a UART. I'm not sure what it is, but it's not a UART, though it can function as one, with some limitations, perhaps.

you say, "... there are quite a few derivatives with a dedicated baud rate generator. Also many SILabs derivatives (e.g. f12x/f13x) allow a separate clock for the baud rate generation. "

That's great, but the original piece of crap is the original Intel version. That's the one designed by someone's son-in-law. The tight coupling between system clock, counter-timer, and serial I/O are the thing that renders it so.

The 8051 was, aside from its serial I/O and counter/timers, probably the first really well-thought-out microCONTROLLER. It has really useful instructions, addressing modes, and other firmware-exploitable features. For the few resources that came in the original, it was really powerful. Sadly, the one set of things, those ones you really love, that hobbled it at the outset have remained. What I'd like to do is to remove the hobbles. In ther process, perhaps I can make the thing able to do things it presently can't.

As for that ultra-high RS-485 ASYNC at 460bps you're wasting several bits per character in ASYNC format. Wouldn't you rather have all the bandwidth at your disposal? If you use 11 bits per character, your RS-485 ASYNC at 460bps becomes more or less equivalent to 300 baud synchronous, doesn't it? (I suspect you should have proofread what you posted.) Wasting bandwidth is an artifact of our society. It's like the packaging on your breakfast cereal. It costs 80% of what you pay, but you can afford to throw it away.

RE