Subject: strange things to do with your IBM System/3x

From: Eric Smith <eric_at_brouhaha.com>
Date: Sat Jan 25 20:06:00 2003

Bill Allen Jr wrote:
> all the ibm languages were available for the sys/34
> and i have all of them basic,cobol,rpgii and fortran.

Somewhat of a brash statement, as PL/I and APL were also
"IBM languages". :-)

> the two sys34's i have both have bad control storage
> cards - a typical problem with 5 years in cold
> storage.

That's a shame! I've got the same problem with one of my
IBM 5100 Portable Computers; the one with both BASIC and APL
fails the non-executable ROS self test. My other unit
only has BASIC.

The 5100 contains three kinds of ROS (Read Only Storage, IBM's
term for ROM). There is a very small high-speed bipolar ROS on
the PALM processor card, containing its microcode. There is about
16K words (BASIC-only) or 32K words (APL-only or BASIC & APL) of
"executable ROS" which is directly mapped into the PALM address space
and contains the system startup, self-test, and I/O code, and the
interpreters for the virtual machines.

Then there is a larger "non-executable ROS" which contains the
virtual machine code for the BASIC and/or APL interpreters. This
is stored in IBM 48 Kbit MOS ROM chips (6K*8). There is no parity
on the non-executable ROS, but the self-test verifies the checksum
of each chip.

Bipolar PROMs suffer from dendritic regrowth of the blown fuses, and
EPROMs, EEPROMs, and flash memory suffer from the charge on the floating
gate eventually leaking away. But I'm not sure what the common failure
mechanisms for NMOS mask-programmed ROM might be.

Sigh.
Received on Sat Jan 25 2003 - 20:06:00 GMT