15 Sep
2001
15 Sep
'01
2:30 a.m.
Tony Duell wrote:
<snip>
In the case you mention, your replacing a TTL device with an FPGA that
currently
uses TTL level I/O.
But most of my classic hardware is pre-TTL and no such logic level compatible
programmable logic exists.
So when a CTL chips fails, I must resort to canabilisim. Eventually it will
no longer be practical to keep this ancient (HP) hardware running. I only
hope that this can be delayed 50 years or so.
But in time, the last examples of the hardware
will cease to run, and
parts will
be long gone. As some (distant point) the torch must pass to emulation,
I am not convinced. It depends on what 'parts' you mean, but if you work
at the level of individual components, then there are probably ways to
make alternatives and will be for many years to come. Consider the AM29xx
bit-slice chips, now moderately hard to find. It's not beyond a seriously
dedicated entusiast to program an FPGA to replace one of those and make a
kludgeboard so it will plug into the original socket. Unoriginal? Sure.
Would I rather use a real AMD chip? Sure, again. But if that's the only
way to keep the last PERQ running, I would have a go.