[sci.electronics] UART for hacking a serial->Centronics-parallel converter?

dplatt@coherent.com (Dave Platt) (09/20/90)

I'd like to homebrew a protocol converter, to connect between a
Macintosh II and a printer with a standard Centronics parallel interface
(an HP DeskJet 500).  I believe that I can do this with an off-the-shelf
UART, a clock circuit, and a handful of glue chips;  I'm looking for
advice as to which UARTs would be well-suited to the application.

What I'd like:

-  Capable of high speed operation (57.6 kbits/second if at all
   possible;  38.4 would be acceptable.  A UART limited to 19.2
   kbits/sec won't cut it).

-  All options jumper-selectable... I don't want to have to put a
   microprocessor in the converter to drive the UART via a bus.

-  At least a few bytes of on-chip buffer memory... the more the
   merrier.

-  Support for hardware flow control when the buffer starts to get full.

-  A simple data-ready/strobe interface on the parallel side.

-  Capable of working with one of those nice Maxim (?) TTL-to-RS232
   interface chips.

-  Reasonably low power, +5 supply only, capable of driving standard TTL
   loads.

-  Available off-the-shelf.

Is there such a beast on the market today?

Yeah, I know I could buy a serial-to-Centronics converter box... but
most of them don't seem to run at above 19200 bits/second, and the
cheapest I've seen costs in the neighborhood of $100.  I'd like
something cheaper and faster, if possible.