Robert Nansel wrote:
I'm looking at what it would take to make a
barebones "museum piece"
bit-serial computer along the lines of an LGP-30 or maybe a Bendix G-1
So I am not the only madman araound here !
Is anyone here sane? I would hope not! I know I'm insane...
And the keep
the project within budget (i.e. none to speak of),
That is, I believe, a recipy
for disaster. Your really need to budget
for having PCB's made if you want some reliable operation of whole cpu.
Why do you need PCBs? This thing sounds slow enough that almost any
construction technique would work (stripboard, for example). And if you
do ened PCBs, they'd only have to be double sided, and those you can make
at home.
[...]
What I haven't been able to get a
handle on is how to make a serviceable magnetic
drum.
A major mechanical challenge, I believe outside most people's
capabilities. Tony is bound to comment on it though !
The comment 'bl**dy difficult' springs to mind...
If you have the heads in contact with the media -- E.g. wrapping
recodinng take round a drup and hainv heads rubbing on it -- you;ll get
wear. Lots of it. It'll work fine for some minuts/hours and then fail.
Rememebr thati n a tape recorder the head doesn't (normally) keep on
going over the same bit of tape).
If you have heads spaced away from the media, then rememebr that the air
gap is critical -- the larger the gap the lower the density of recoridng
yoyu can use. But the gap, in turn, is detemriend by the mechanical
tolerances of the drum and bearings -- and you're going to have problems
from that. I would think any form of ball/roller
bearing is going to have
too mich run out -- even an adjustable tapered toller
bearing. I would
think tapered bush bearings (and used in a good lathe headstock) would
do, but making them -- and adjusting them -- would be 'interesting'.
-tony