16 Oct
2009
16 Oct
'09
7:43 a.m.
On 10/16/09, Rick Bensene <rickb at bensene.com> wrote:
The DF32 carries its own controller in the primary unit - there's room
behind the faceplate for several "system blocks" of 8 single-sided
modules or cable paddle cards. The "master" DF-32 has a full load of
slots and modules to implement the IOTs and the 3-cycle databreak
logic. The expansion drives, also called DS-32, have fewer slots and
modules because it's only the logic necessary to read/write the drive
and talk to the master.
The likely reason they were afraid of power failures is that the heads
do not retract on a DF32. They land right on their tracks. Too many
spin-down/spin-up cycles is bad. The drives, in fact, have the
ability to disengage power to the logic while leaving the motor
spinning - you can do basic maintenance without a complete power-down.
You can hook up just a single DF32 to any negibus machine with just a
large wad of cables - there are a few modules inside the CPU for bus
interfacing and the CPU component of the 3-cycle databreak, but I
think (since I do have this gear) that there's no "controller" between
the first drive and the CPU, unlike, say, an RK8 or TC08.
On 10/14/09, Rick Bensene <rickb at bensene.com> wrote:
> OMSI had a Straight-8, with quite a bit of RAM(er...core something
like
24K?), one of the fixed head disk drives (was it RA32?)...
Probably a DF32.
Yup, that's it...DF32. It was housed in a full-height rack. I recall
the controller filled quite a bit of the rack...seemed the controller
was more complex than the Straight-8. I remember that the folks there
were always very afraid of power failures. I was there once when the
power did bounce a bit, and everyone got very nervous that the drive
might crash. It seemed to make it through OK. Another thing I remember is that one time something
went awry with the
DECTape controller. A DEC FSE came to fix it. Turned out it was
something with the power supply for the controller. The power supply
was replaced, but somehow a connection got swapped somewhere, and when
it was powered up, the controller let out all of its magic smoke. It
took a couple of weeks before the controller was fixed.
Oof! If it was an R-series-based controller (contemporary with the
CPU), the voltages coming off the PSU are +10V and -15V. Swapping
anything around would likely blow a *lot* of stuff up. An M-series
controller (built after, say, 1968), would be +5V and GND for sure,
and probably still the -15V here and there - still not good to mix up.
Somewhere around here I have a DECTape from that
system...I took an
assembly language programming class there and my class project (printing
text banner pages on the Centronix printer) files were on the tape.
That would be an interesting tape to find. Unless it's been stored in
an uninsulated attic or in a barn, I'd expect it could still be read -
they are very robust. I don't presently have a DECtape drive up and
running, but many people on the list do and could help, once you run
across it again.
-ethan