William Donzelli wrote:
That's
true, but there are decades worth of NOS TTL chips on eBay. I've
gotten a bunch of them recently of the types I think I'll ever use. And
really cheap at that.
And after those decades worth are gone, a few more decades worth will turn
up with just a little looking. We will all be dead when the things become
unobtainium.
I am always getting into arguments with folks that are worried that the
1960s transistors used in early minicomputers (PDP-1s, Packard Bell 250,
and the like) are "all gone" or "extremely rare". 'Taint true,
Dooley -
all you have to do is look.
Concerning transistors of the 60s (even Germanium-type PNPs), I wouldn't
have a problem to use modern types
as replacements if the go or no-go of an old system would depend on it.
Discrete electronic components are rarely a
problem, even coils and transformers could be replaced. More difficult
are certain kinds of integrated circuits.
The mentioned 7480 is maybe hard to locate even as a datasheet (good
idea to keep old TI TTL databook, IIRC
also at bitsavers), and even more difficult to locate the chip itself,
but this is a function that would easily fit into a GAL.
A real problem are species what can't be simply replaced by some GAL,
namely open collector components or "extendable"
gates like the 7450. Replacing a single chip of such a kind usually
means replacing its environment as well, e.g. that 7450 would
then require to pull the extension gate (was it 7460, my memory is
failing) as well.
In the area of unobtainium, however, you usually don't find SSI or MSI
circuits, but rather nasty special analog circuits and
early LSI MOS circuits. Often a modern uC could emulate the latter, but
this often needs significant tinkering for each problem,
and an many cases also rewiring its environment.
Surely the question original replacement vs. reengineered functionality
has some religious aspect, but if it is the question of whether
an old system should be left inoperable in a museum in contrast to
actually switching its blinkinligths I favor reengineering.
Regards
Holger