AW: Matrix board! (a CMOS one)

[Edu Silva]

De: Haible Juergen <Juergen.Haible at nbgm.siemens.de>

> That alone would not help at all. If it's just switches, it would more or
> less short whole columns and rows together.
(snip)
> ...What you need is a resistive connection for each matrix element, with
> the input
> rows being buffered and the output columns being true summing nodes.
> It would not be completely out of reach to integrate such a scheme
> (transmission
> gates are not perfect shorts), but I have my doubts that this implemented
in
> these arrays.

I'm aware. I've designed and built matrix mixers/switchers for PA
monitoring and paging systems.

The first thing I did when I got the sheet on my hands was taking a look at
the block diagram, and didn't see buffers or anything like. The drawing,
however, seems to be just illustrative, as all the logic section is
represented by only two blocks: one is the decoder and another, the
latches. DATA, STROBE, RESET and CE pins are not shown, but they exist.
IMO, buffers and summers were ommited just for the sake of clarity.

Nowhere in the text is said that we can't mix the inputs. I think that if
folks at Harris have made this thing able to perform any combination of
switches ON (as it seems), they must provide means to do it with safety.
OTOH, if mixing is impossible, the logic section must prevent it, but it
doesn't (well, that's the way I do when designing such things...).

Thanks for the interest, anyway  :). I'll ask the guys at Harris and keep
the list posted.

Edu

_____________________
ES2 Audio - Som & Luz
Salvador-Bahia-Brasil