Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!ames!oliveb!amdahl!esf00
From: esf00@amdahl.amdahl.com (Elliott S. Frank)
Newsgroups: comp.misc
Subject: Re: What the world needs now [ is an exploding computer ]
Message-ID: <7696@amdahl.amdahl.com>
Date: Mon, 1-Jun-87 13:12:27 EDT
Article-I.D.: amdahl.7696
Posted: Mon Jun  1 13:12:27 1987
Date-Received: Wed, 3-Jun-87 01:19:41 EDT
References: <12067@topaz.rutgers.edu> <910@killer.UUCP> <15@gordon.UUCP> <2725@phri.UUCP> <8252@bu-cs.BU.EDU>
Reply-To: esf00@amdahl.UUCP (Elliott S. Frank)
Organization: Society for Infinitesimal Procrastination
Lines: 44

In article <8252@bu-cs.BU.EDU> bzs@bu-cs.BU.EDU (Barry Shein) writes:
>
>I can't verify it but [. . .] the heads
>just banged back and forth till they blew a disk drive
>
>Another similar story referred to [...] tape drives
>which, when used by programs which did direct access to tapes,
>eventually stretched and broke the tapes

The "washing machine test" stories may refer to the IBM 2311, or it's
predecessor, the 1311.  The 2311 put 15 megabytes on a stack of 5, 14"
platters (the 1311 put 7.5 Mb in the same area), using a combination of
(oil-filled) hydraulics and voice coil technology to position the heads.
Total mass of the head and arm assembly must have been in the hundreds
of grams range (any designers out there who can now say what the mass
was?) For ease of installation, the drives had 5" rubber wheels, which
were supposed to be jacked off the floor when the drive was positioned.
When the 360/75 was installed at Columbia (late '67), one drive was not
positioned on its jacks.  None of the installation tests caught the
problem -- a disk sort that put multiple workareas on the misinstalled
drive did.  The drive didn't get a chance to go very far -- the disk
cables to the controller were IBM systems cables (24 minature coax,
about 1 1/2" thick) that did not have any slack.  The largest
damage was to the ego of the FE who didn't check his checklist.

The tape drive stories may be due to the IBM 729 drive, originally
designed for 200 bpi, and then upgraded to 556bpi and 800 bpi densities.
If you did a fast forward and then issued another command before the
fast forward completed, you could get the drive to do interesting things.
One trick used by DCS was a fast forward (to the end of the tape and wait
for it to complete), fast forward, rewind sequence that allowed it to
read past a BOT/EOT marker. That trick allowed you to put both the IBSYS
and FMS operating systems on the same single reel of tape, with a BOT/EOT
marker in the middle.  If you used a single FF to get to the end of the
tape and then asked for another operation, you had a problem, as
the controller did not implement the equations of motion.

-- 

Elliott S Frank    ...!{hplabs,ames,seismo,sun}!amdahl!esf00     (408) 746-6384

[the above opinions are strictly mine, if anyone's.]
[the above signature may or may not be repeated, depending upon some
inscrutable property of the mailer-of-the-week.]