On 3/2/12 3:04 AM, Christian Corti wrote:
The DiscFerret won't help with this drum either.
Speaking of experience (having two working LGP30s), you don't really need to digitize
the drum if the machine (or more exactly, the matrix drivers and
read FFs) are working, just hook up your logic analyser or whatever to the read FF and
select the desired track. Your read clock is the main clock (137 kHz for the European
version). Since there's no
index pulse you need to read the C register together with the two clock tracks S2*/S3 to
get the number of the next sector. The word mark is the sign bit. ... But you really only
need to do that if
you absolutely can't get your LGP30 working.
When we got our first LGP30, before the machine itself was operational, we had dumped the
drum with a Dolch logic analyser and found a mostly empty drum with some code fragment
somewhere in the
middle. We used some simple TL084 opamps to digitize the signal and feed it into the
pod.
Many years later we discovered that this is what is left after running the LGP30
acceptance test.
Providing detailed information on this would be helpful in the future if someone needed to
build a non-mechanical replacement for the drum.
I know of at least one Bendix G-15 that could probably be checked out if a replacement for
the drum could be put together.
This information really should be collected for any machine where the timing information
that drives the computer is on rotating media
(LGP-30, G-15, IBM 650, Recomp-II, D17B)