On Jan 4, 11:08*am, Tim Wescott wrote:
...one, what would I do if I were building pure one-off (I'd
use one of the old PC's I have lying around, with a digital I/O card);
....

I've built or maintained a lot of one-off test systems of similar
complexity, which usually were assembled around an old desktop or
laptop with I/O ranging from a small wirewrap board on the printer
connector to a rack of GPIB instruments.

The PC brings a nice built-in development environment and the keyboard
and display, all very useful when the application needs considerable
monitoring and adjustment or data logging.

Their power consumption isn't too bad. I've measured around 30 - 40
Watts for PII desktops, half that for laptops running without their
dead batteries on the AC adapter. The 400 MHz Armada laptop I use for
datalogging draws 14W at idle.

If you decide to put your circuit on an internal wirewrap prototyping
card the bus interface can be quite simple. One that I did for the VME
bus consisted of a very wide gate combining the addresses and read/
write commands that clocked a single D flop to send the acknowledge.
My ISA and Mac NuBus interfaces weren't much more complicated than
that. I think you can find proto boards with a full bus interface
installed, though you can get by with mapping the board into IO
address space. ISA is easier than PCI.
http://www.epanorama.net/documents/pc/isabus.html

Laptops are a little harder to talk to. Some PICs support USB or
serial data. The printer port is really quite useful for both serial
and parallel I/O if you run DOS, but Windows polls it and interferes
with the control register.
http://www.epanorama.net/circuits/pa...tml#circuithow
http://www.doc.ic.ac.uk/~ih/doc/par/
By decoding two control lines you can have 32 read/write bits in four
registers, plus Read and Write strobes. You can sneak two more control
bits out of a serial port.
http://www.aggsoft.com/rs232-pinout-...and-signal.htm

jsw