I don't
think anyone else managed a port to that particular CPU
National ported 4.1 to it. Their version is called Genix.
Tektronix' first entry into the UNIX "deskside" workstation market, the
6205, ran a modified version of Berkeley 4.2 called UTek on the 32032
CPU. The next entry, a smaller, desktop workstation, the 6130, used the
32016. It was a little smaller than an original IBM PC (except a bit
deeper), and had a backplane bus that explansion (RAM, SCSI, Serial I/O
(2 ports per card), Parallel I/O) could plug into. It had an onboard 10
megabit Ethernet controller (AUI connection) with a TCP/IP stack, two
built-in serial ports, and a GPIB port. It had a built in 5 1/4" floppy
drive which was used to load miniroot, and install UTek (a terious and
slow process with lots of floppies). The machine used an internal
ST-412 compatible hard drive, originally Micropolis 20MB drives. Later,
Maxtor 80MB and 140MB drives were offered for the 6130.
A later version of the 6130, virtually identical except for ST-506 was
abandoned in favor of an internal SCSI controller, and internal SCSI
drives became available. This was called the 4132. It provided a SCSI
port on the back for adding additional SCSI. Instead of a built-in
floppy drive, the machine had a QIC-24 5 1/4" half-height 1/4" cartridge
(DC-300) tape drive, which had an Adaptec SCSI to QIC-24 converter board
attached. This made it MUCH faster for loading up the OS. Internal
drives up to 320MB were offered, IIRC. The 4132 was also a little
faster than the 6130, due mostly to an improved I/O subsystem. I
believe the clock rate of the 32016 was the same. The 4132 did "feel"
faster than the 6130, though.
UTek was pretty much 4.2bsd compatible. Tek tweeked a few things (for
better or worse). However, initially, UTek's had some TCP/IP networking
issues which were troublesome -- the network stack didn't understand
netmasks properly, making it difficult to attach these machines to
subnetted networks.
UTek had a number of applications packages, including ANSI BASIC
(interpreter), a full C programming environment, a spreadsheet
application called 20/20, (all text based), and a few others that I
can't recall off the top of my head. A graphics processor was designed,
that piggybacked on the CPU board, that allowed X-windows to run (very
slowly) on an attached color video monitor, with a special keyboard and
mouse. These were very rare, though. Most 6130's and 4132's hooked up
to "dumb terminals", or Tek's line of terminals with various levels of
intelligence, including the DVST (Direct View Storage Tube) terminals,
and later raster-terminals that ranged from glass-TTYs to complex color
graphics terminals, all over RS-232 connections at up to 9600 baud.
I have a 6130 that needs a little TLC, but I'm sure could be made to
work. I also have a fully funtional 4132.
I also have a lot of expansion boards (RAM, Serial I/O, SCSI), which
work both in the 6132 and 4130, and a few spare power supply units, and
even some spare ST-506 Micropolis drives (20 Meg. Drives that have been
"tweeked" (long story) to provide 30MB), and miscellaneous other parts.
Haven't had a chance to tinker with any of them for a long time, so
there'd probably be teething pains in bringing the 4132 back online.
Reparing the 6130 is probably something that I won't get to for years.
The biggest problem is that the Wangtek QIC-24 tape drives have a
tendency for the "rubber" tape drive capstan wheel to turn to goo,
rendering the drive useless (and will destory tapes instantly if you put
a tape in a drive with such a failure). I have a number of the drives,
and all of them have gooey capstan wheels, except the one that is in the
4132 at the moment, which I actually "rebuilt" some time ago when I was
able to find a couple of "dead" drives that had good capstan wheels on
eBay that I got for very cheap.
Today, these are very hard to find.
Later, Tektronix abandoned the National 320xx line, and came out with
some UNIX workstations based on the Motorola 68020. UTek was tweeked to
make it more SysV compatible, and these machines had built-in
framebuffers and could run X relatively decently. They could handle
more RAM, and had higher I/O bandwidth.
As Tektronix' last hurrah in the UNIX workstation marketplace, they
moved to the Motorola 88K CPU, in boxes called the "XD88" series. These
machines were actually quite ahead of their time -- they were quite
fast, had tremendous graphics capabilities, and lots of expansion
options. The graphics processor was designed by Tektronix, and I
believe (but am not totally sure), that it used another Motorola 88K, as
well as some ECL bit-slice technology and very high-speed static RAM for
the framebuffer. The board was an intelligent graphics processor,
capable of rendering objects posted in display lists, such as circles,
arcs, squares, rectangles, triangles, and various other structures
without main CPU intervention other than just populating the display
list. Fills, bitblts, and other functions were also performed by the
graphics processor.
Sadly, Tektronix really wasn't a computer company. They didn't try
going after Sun, Apollo, or other UNIX / UNIX-like system vendors of the
time. Tek's marketing people, who were more focused on test equipment,
targeted the earlier machines as hosts for automated test systems using
the GPIB interface to talk to Tektronix instrumentation.
This way, they could "sell" the workstation as a controller for more Tek
instrumentation, which is what the sales force was used to selling.
Later, the XD88's were targeted at MechCAD and EDA workstations, but by
then, there were a lot of players that were well-established in this
space. Because of the Motorol 88K processor, applications had to be
found and ported, which proved problematic, and expensive.
Tek had some workstations which were better than many of the competitors
of the time. The 4132 was faster than the early Sun 3 machines,
especially the "big pizza box" units like the 3/50, and though the 4132
didn't generally have X, as a command-like development machine, it was
very useful. The C compiler was quite fast, and generated efficient
code. A system with 3 Megabytes of RAM ran very comfortably with very
little swapping unless you were really beating the heck out of it. It
handled multi-users quite nicely. I believe the 4132 could be expanded
to up to 7 Megabytes of RAM. Had Tektronix properly realized the
marketplace, and targeted areas other than test & measurement with the
machines, they may have been successful with them. But, Tek had a very
long history of being nothing but a company that sold Test & Measurement
equipment, and it was a difficult sell to convince the sales force, who
all did VERY WELL selling Tek's oscilloscopes and other test equipment,
that selling computer was going to be as lucrative. As such, the
machines never really caught on much.
I know a lot about these machines because I worked for Tektronix from
mid-1977 until late 1990. I actually built the 6130 that I have by
buying the parts from Engineering Stock (employees could by parts for
anything that Tek made through Engineering Stock for cost+10%) and put
it all together myself as a "kit". It was quite an adventure.
Great thing was, it ran straight away when I powered it up. It was
quite cool to have a "real" UNIX workstation at a time when other people
were playing with Apple II's, Commodore & Atari machines, and S100
boxes. I actually put it online as a public-access UNIX system for a
time, and also wrote (in ANSI BASIC) a computer bulletin board system
that I ram for a time on the machine. Those were the days when I
actually had spare time. Those days are LONG gone.
Rick Bensene
The Old Calculator Museum
http://oldcalculatormuseum.com