2 Apr
2002
2 Apr
'02
3:40 a.m.
Richard Erlacher wrote:
What I've encountered more and more is the inconvenience associated with
trying to use current-generation components in wire-wrap. I'm always having
to build adapters that make a DIP out of a TSSOP or the like. The adapters in
a prototype often exceed the cost of a PCB. It's much easier to build small
boards, about the size of a typical playing card, and that's enough space to
house a microcomputer of reasonable capability, comparable with any "classic"
8-bitter of the '70's and '80's, including its I/O, memory, video, and
mass
storage interfaces. After I get the current task off the table, I'll take a
closer look, but it seems to me that it's easier to put a system on a
playing-card sized board, and cheaper too, using current technology, than to
recreate the old stuff using parts that are increasingly scarce and boards
that are needlessly large and costly.
I would suspect that one would need a larger size PCB because that size
of PCB can't hold things bigger than 56? pins. A reasonable CPU just
fits in a 84 pin PLCC. Somebody (link not handy) makes a PLCC to
breadboard adapter on the net, so you can use a breadboard for the newer
(small) chips. It is the speed of the old I/O chips that are a problem.
A 1.5 Mhz+ chip is high speed and all the good chips like floppy disc
controllers are all in a mother board chip set for a PC. Mind you 8 bit
CPU's are limited to 64kb as we all know is way to small for something
other than a minimal PC.
My own CPU design (in a FPGA) is limited to 512kb of addressing range if
I want to emulate a fictional CPU in a 40 pin dip. While I use 32kx8
static ram and few special control signals and a UART in the FPGA most
of my design is what a possible 1980's cpu could have been like: Limited
front panel, Serial I/O, Dynamic Ram,Mother board with CPU.
--
Ben Franchuk - Dawn * 12/24 bit cpu *
www.jetnet.ab.ca/users/bfranchuk/index.html