From: HOTZE <photze_at_batelco.com.bh>
<Hello... some time ago, there was talk of building a computer, and now I
<think that I've got a (bad, possibly) idea. In the earlier half of this
Not a bad idea but, it appears you have no technical concept of the extent
of it.
<century, transistors weren't avaible... vaccum tubes... huge ones, but
<now, the transistor has made small ones possible. My point: If we were
<to take a tubed design, and re-build it with transistors, we could
<probably make it a decent size.
< So, what da ya think?
The operating characteristics of tube and transistors are far enough apart
that circuit techniques applied to one donot apply well to the other.
The redesign would not be tivial
For most of the tube designs drum memory and mercury filled acoustic delay
lines were memory. a few used williams tubes. You would find that
difficult to duplicate.
To further make a point most of the early transistor designs were
evlotionary results of tube designs.
One of the first transistor designs was TX1 and TX2(early 50s, MIT/lincoln
labs) and theywere not small. Later ones in the 60s were PDP-1, Perkin
Elmer, CDC, to name a few and these were large as large machines(fast, big)
were the goal. The first small machines were the LINC and PDP-8.
Figure 5-10 thousand transistors and thousands of diodes, resistors,
capacitors. A foot print for a typical transistor -8 is aroung 20 square
feet plus access space.
The closest you could come and expect to complete on a singular basis would
be a to copy the archectecture of an older machine using ICs. It wouldn't
be the real thing, original peripherals would be hard to come by and there
may be a base of software if you elect to copy something like a PDP-8.
I'd do it as an engineering exercise and beacuse I happen to like certain
old archetectures. The result would not be a classic by any means.
Allison
Received on Thu Nov 20 1997 - 10:53:06 GMT