14 Nov
2006
14 Nov
'06
11:12 p.m.
Josh Dersch wrote:
The RD53 drive in my 11/73 has finally gone to silicon
heaven, after a
brief resurrection. Given the reputation these drives seem to have,
I'm not too surprised. So a few nights ago I dug through my stacks of
ancient hard drives and found three possible candidates for replacement:
RD53... Micropolis 1325 I believe. I ran one of these (I believe it was
an RD53
actually... not just a generic 1325) in a PC years ago.
It one day started not wanting to initialize... startup... think think,
chunk chunk,
spin down....
With nothing to lose, I began to experiment...
What I found was that it couldn't apparently locate track 0. If the HDA
os opened,
it has two rubber? bumpers for the extreme ends of head motion. One for
track 0
and one for last track ?
They look something like:
----
| \
|____ /
with an adjustable slide 'hole' in the middle.
What I'm guessing went wrong is that the material hardens and/or shrinks
with age.
By carefully tweaking the adjustment on this item to allow a mere few
thousands of
an inch inward... the drive started initializing fine every time.
I did not use the drive much after that, so how drive longevity would be
effected by
the introduction of contaminants into the HDA assembly I don't know (I
of course
don't have a clean room :-) ).
However, if you have a fairly clean space, and can work carefully as
such not to
stir up contaminants, you may be able to rescue your RD53.
Another problem I read of someone having a problem with an RD53 had to do
with the brake solenoid assembly needing adjusting as the brake pad was
dragging
(it is located under the circuit board on the bottom as I recall... no
inside the HDA).
-- Curt
- A Microscribe 6085, which seems to have the same drive geometry as
an RD53
- Two Seagate ST-251s, which seem to be identical to an RD32
All of these are in "unknown" condition (I haven't used them in
years), but spin up without making any evil noises and are recognized
by the RQDX3 controller. To the best of my recollection, all three
were formerly used in PCs.
I've used zrqch0 (booted via vtserver) to format all three of these
disks. In all three cases the formatting appears to be successful --
the drives format for a few minutes, then go through three
verification stages for another 20-30 minutes. During this time I can
hear the heads being stepped, and the drive access light blinks. At
the end of the format no errors are reported.
I'll note that in order to do the format on these drives, I had to go
through the process manually (as opposed to the "AUTOFORMAT" mode of
zrqch0) -- for the Microscribe this was because zrqch0 didn't
recognize the drive (showed up as "unknown"), for the ST-251s this was
because the fault tables could not be read off the drives. I suspect
this may have something to do with the issue I'm seeing, but of course
this is pure speculation. I've tried various permutations of the
formatting options, but I've had no joy there.
After formatting the drives, no other utilities seem to be able to
make use any of the drives -- running the disk exerciser (zrqah0)
fails after the first test step (and after a very small amount of disk
activity) with an error similar to the below (this varies slightly
based on the test being run):
FATAL I/O ERROR
* DISK 0 WENT OFFLINE
* SUB_CODE: NO VOLUME MOUNTED OR DRIVE DISABLED BY SWITCH
* COMMAND: READ_COMPARE
Interestingly, 2.11BSD's 'disklabel' is able to partition and write
the label to the disk. Mkfs, however, fails with:
ra(0,0,0) error op=A2 sts=23
write error 3
Exit called
I've tried this with two different RQDX3 controllers with the same
results (and these controllers worked fine with my RD53 until it
failed). Is there a magic step I'm missing? Is what I'm attempting
to do (use non-DEC drives) not possible? Any ideas or suggestions?
Thanks for any help here...
Josh