[Going from memory]
> The Fairchild FST-2 computer was used to control a family of
> Semiconductor ( chip ) testers I guess from sometime in the mid 1970s to
> around the mid 1980s. The family of testers were branded as "Sentry",
> "Sentinel" and "Series10" ( maybe some others ).
The Sentry 20 and 21 were the last testers that used the FST-2. I
programmed them for a few years. The Sentry 15, 50, and 90 were
completely different systems (and developed later than the Sentinel
and Sentry 10).
> The FST-2 was a 24-bit machine with its memory interleaved between odd
> and even addreses. When the CPU was reading/writing to an odd address, a
> peripheral could address an even address and vice-versa using DMA. Early
> machines were booted from tape, but newer machines had a PROM boot board
> which allowed booting from 8" floppy, hard-disk, tape ( and possibly a
> network. ) Early machines had 25 bit wide memory with a parity bit,
> while the later machines had 30 bit wide memory which provided error
> detection and correction. Control of the tester was via two busses known
> as the long-register bus and the short-register bus.
I seem to remember that network booting wasn't available, but you could
boot a minimal system from tape and load the rest over the network. We
had RAM disks in our systems that we booted from, so the only time we
had to boot from tape was when we had to power the system down or change
from M3 to the diags. When we had FASTNET added we changed to load most
of the files over the network for either M3 or the diags and that sped
things up significantly.
> If you come across an FST-2 which can be powered up, set the console
> switches ( piano keys ) to 06760000 in octal and press STOP, RESET, LDP,
> LDC and START. The machine should waken up. There are a few stand-alone
> FST-2s kicking around without any tester hardware. These were used as a
> FACTOR compiler co-processor for a development system which was based on
> the HP1000. Most FST-2s will be part of a chip tester. If the chip
> tester is still working, the whole system is still quite valuable. Even
> as a source of spares.
The HP was phased out later in favor of a VAX 11/750 (and later a MicroVAX
II). The testers could load test files over serial from the HP and either
serial or FASTNET with the VAX systems. We eventually upgraded from a MVII
to a MicroVAX 3900 (the fastest system FASTNET could run from as it used a
DRV-11W and the driver wouldn't work on a VAX 4000 system due to the changes
in the Q-BUS timing).
> Although the tester has a lot of different power supplies the FST-2
> probably ony needs 5.0V and the RS-232 voltages to get running.
>
> The FST-2 ran an OS named M3 ( "M Cubed" ) and was programmed using a
> FORTRAN-like language named FACTOR, ( Fairchild Algorithmic Compiler
> Tester ORiented. ) Most Sentry testers could test digital chips with 60
> pins up to 10 MHz. Some later Sentrys could test up to 120 pins at 20
> MHz. I remember FST-2s being very slow to compile FACTOR programs. Most
> users ended up compiling their program and test patterns (
> Vectors/Truth-Tables ) on a VAX.
Actually M3 was a follow-on to an earlier OS (I think it was called Flopsy-
Dopsy, but I can't remember the real name). M3 was developed by IBM and
sold to Fairchild.
The Sentry 20 and 21 testers could test at 20MHz. The biggest difference
between the two systems was the timing accuracy. You could also get the
systems in a high-speed or high-voltage configuration. The HV configuration
allowed higher voltages, but lower timing accuracy (slower rise and fall
times on the timing generators).
FACTOR was an interesting language. You could also write programs in
assembly language, and also write programs to execute in local memory
in machine language (which is not local memory on the FST-2, but local
memory in the test head).
> Fairchild ( and later Schlumberger ) provided really good in-depth
> training and documentation for the CPU and the testers for hardware
> maintenance and programming. The manuals which were up for grabs were
> either manuals/schematics which were shipped with a tester, or training
> manuals which some engineer picked up in San Jose or Munich.
I think I may still have my manuals somewhere, but I may have thrown
them out.
--
Eric Dittman
dittman_at_dittman.net
Check out the DEC Enthusiasts Club at http://www.dittman.net/
Received on Mon Dec 24 2001 - 23:17:09 GMT