Re: wonderboard

>
> From: "Harrison, Doug" <dharrison@...>
>
> Couple months, maybe. I will need some help with the layout and component
> selection. Will probably produce a few beta versions to make sure we get it
> right. Anybody object to using a wall wart to feed the 5vdc supply?
>

In no real order:

My offer on the parts still stands.

Just to give an idea of the number of some of the parts I have,
in case someone if thinking I have small box of junk...
74-ALS-LS-F-244 Tri-State Octal Line Drivers, Rcvrs.
I have 1,584.

Almost all the parts I have are LS type TTL parts.
It seems they have the advantage as far as drive and sink currents.
They will source 24 ma, this should be able to directly drive many
LEDs and Optoisolators.

The issue of long signal line runs and noise would better be addressed
by differential line drivers than 120VAC, IMHO.
I have a bunch of RS-422 line drivers and rcvrs as well.

I have a bunch of PCB mount relays as well, I'm sure of the rating,
I think they are single pole double throw, they are about a 1/2" cube.

Here is MY take on the interface.
To give the most flexibility and largest address space to the interface
We could make access a 1,2 or 3 step process as necessary.

First Write an address and command string, the only command I can
think of at the moment is read or write.
Something like:
0 1 1 0 1 0 0 1
? ^ +-adderess+
|
0 for read 1 for write
The lower 6 bits are the address for 64 possible locations.
This would allow 512 bits of input/output.

Watch for a ready signal on one input line(If Appropriate)
Read or Write the data.

I would suggest presuming a bi-directional interface and not do the
4 bit crap.
The PS/2 type bi-directional interface has ONLY been around for 13 years.
It SHOULD be supported even in junker type $5.00 hardware.

Use the additional I/O bits on the port for stuff like the ready flag
mentioned above.
The only time I can think of right off the ready bit would be used
would be handshaking for talking to an external CPU or something,
A delay for an A/D converter to do a conversion after a read request,
and to sync up to another external action.

Otherwise this thing would run in a latched state and you could set
up some of the TTL buffers, to pass data straight thru and would
have any input changes instantly show up at the printer port until
you set a new address.

Such a system could easily be built with TTL parts, and with a simple
design based on latches it would be easy to understand and troubleshoot.
A lot simpler than the 8255 btw.

Adding an Eprom directly wired to a pair of 7 segment displays could
decode the address at any given instant and show the number of the
currently active port in decimal digits.
Then again maybe it would change too fast to bother with...

BG Micro has 7500 V optoisolators for 17 Cents.

Marc