4 Mar
2012
4 Mar
'12
12:34 p.m.
On Sat, 3 Mar 2012, jim s wrote:
I would love to know what steps were used to read off
a drum, if the unit had
to be functional to operate the drum, what was modified, how was the drum
spun up, what external circuitry was used to obtain the data. I think there
is one LGP-30 drum in our posession and two LGP-30's, one functional, one
maybe not so which all have drums.
First, I my opinion it is not so interesting in preserving the contents of
an LGP-30 drum, simply because there most probably will be either nothing
(e.g. after running the drum or acceptance test), or just 10.4 at the
beginning, then e.g. the ACT runtime system, variables and temporary words
(at track 63). There may be some text strings of a binary program, but
there won't be any sources.
But anyways, if you are looking at the drum, you'll cleary see what is
needed to read it: run it (motor), take the clock from the clock track
(clearly visible, its output is the coax cable), and the three timing
tracks. Then you can select any data track you like, the diode selection
matrix is almost self-explaining. We did all this even before we received
any documents or schematics of the machine. It really is very trivial.
We didn't modify anything, the drum was spun up by pluggin in the power
cable into the outlet (for the first try, we attached a power cable
directly to the drum motor; normally it goes to the relay box where
power is applyed after 50 seconds).
Before I'd embark on the task I'd love a
detailed explanation of what you or
your group did to do retrieve the data and store it, passively or if you had
to restore things, what and how.
There's not much to explain. The LGP-30 is a very simple and solid
machine. Clean all contacts (including tube sockets), have a look at the
power supply, relay box and magnetic constanter, look for broken wires,
disconnect the drum motor and fire up the machine. If it will pass the
three power-on phases you'll have an (at least almost) working machine.
Hook up the drum and start examining the contents. You may have a
defective diode on the logic board or on one plugin board, but that won't
destroy anything. Take a scope, trace any fault, repair it, et voila.
There have been accounts here recently by people
detailing what they did in
detail which would be helpful for the restorers of such hardware, and that
you digitized your drum is very interesting.
Is it? I find that very boring, especially since the task was unnecessary
because the machine worked almost right away (after 20 years of storage).
I think that we used the TL084 to digitize the read signal and feed the
outputs to the set and reset inputs of an RS-FF (e.g. SN7474 or SN7400).
That's it; it's essentially what the matrix drivers and read flip-flops do
in the machine. You can watch the signal on a scope or logic analyser and
decide whether there is something you want to preserve or not. In our case
the drum contents of both machines was very unimpressive.
The construction of a replacement would be aided in
the insight gained in
interfacing to the existing drum.
I'd probably use a microcontroller to generate the timing signals and some
sort of serdes with SRAM for the data. I've heard of at least one LGP-30
that had been upgraded this way.
But the details are described well enough in the serivce manual.
Christian
PS:
Apparently the English maintenance manual is not available, at least I
can't find it. The link to the German one has been posted several times
here, it is
ftp://ftp.informatik.uni-stuttgart.de/pub/cm/lgp30/docs/WHB.pdf