C.H. Burrow wrote:

> The last time I played with eproms in this way I found the outputs
> quite glitchy, the effects of which become especialy noticable at speed.

 Ah, but the LED driver application would appear to represent anything
BUT
speed!  At a multiplex rate of 100Hz, I am sure that the odd 50ns glitch
would hardly be visible!

> If you decide to use buffers on the outputs, perhaps latched ones  that
> latch when the eprom outputs have stabilised may be a good idea.

  The use of latches might in fact be counter-productive, as you then
need
to provide a mechanism to delay the clocking of the latch from the
update
of the addresses.  If you are going to provide such a delayed strobe,
you
can just as well use it to enable the device which is using the EPROM
pattern, disabling it until then to avoid false patterns.

  Cheers,
        Paul B.