Tektronix 8501 Data Management Unit

From: Rick Bensene <rickb_at_bensene.com>
Date: Mon Nov 22 19:25:22 1999

Tony wrote:

Re: Tektronix 85xx development systems

> Yes, there are various versions of this, mostly somewhat
> similar. I think
> one of mine claims to be an 8550 (and I think that's the name
> for the 2
> boxes together).
>

Could be right...maybe it was 8550 not 8560, tough to remember.

> >
> > The machine I had was in a similar case to the 8501, but inside was
> > a LSI 11 CPU board (original DEC, as I remember), and non
> Q-Bus backplane
> > that Tek devised that could accept Tek-made memory and
> peripheral boards.
>
> Yep. The bus is electrically very close to Q-bus, but of course the
> connectors are totally different..

I remember being frustrated once finding a Q-bus memory card at the
country store in the hopes of using it on the system, and then having
someone there tell me that the 8560 was not Q-bus compatible.
Didn't know if it was electrical or mechanical incompatibilities.

>
> > The machine that I had had a Micropolis 8" harddisk (can't remember
> > the model), and a Qume DT-8 8" floppy drive. The Micropolis
> > drive sounded like a jet airplane when it spun up, along with the
> > really noisy fans in the cabinet itself. It was almost
> uncomfortably
> > loud.
> >
> > I think the hard disk held a total of 35 Megabytes. The
> machine did run
>
> Exactly. A Micropolis 1203. I know it reasonably well, having
> repaired
> them from time to time.
>
I was pretty enamored with the drives at the time...they did seem
to be really reliable.

I remember Tek's first use of these drives. Tektronix built a
graphics workstation VERY early on...long before Sun existed...perhaps
just after Xerox and their early graphical machines. There were
visions of selling these things, but Tek's management squashed the
idea. The machine was called Magnolia. It had a large (19"?)
monochrome (white phosphor), a fairly high resolution monochrome
frame buffer (perhaps something like 1024x768), a keyboard & mouse, and ran
a windowed operating system
that I think used Smalltalk as the development environment, and some
form of Unix kernel as the OS. The Micropolis drive was used in the
machine.

Don't know what the CPU was...I think it was some custom bit-slice
based machine. I recall that it was quite fast.

The machine was built on a pedastal, with the HD in the bottom of the
pedastal, and the CPU, memory and framebuffer higher in the pedastal,
and the monitor & keyboard up top. Similar to the pedastal-mounted
Tek DVST graphics terminals like the 4014. Anyway, they number of
in-house prototypes that they built had major problems with the HD's
crashing, because people had a habit of swinging their feet while sitting
at the machine, repeatedly thumping the pedastal...and the vibration
would cause HDA crashes. Apparently this was further aggravated by the
fact that the HD's were rather picky about temperature, and the pedastal
didn't afford very good cooling for the drive...overheating was a problem.
I saw one of these beasts in operation, probably sometime around 1976.
It was truly amazing at the time! Had Tek decided to market these machines,
and done a good job of it, they could have beat folks like
PERQ, Sun, and others to the 'windowed Unix workstation' market.

> > Tektronix-munged version of V7 Unix called TNIX. The intent of the
>
> I thought TNIX needed an 11/23 CPU card (which was normal on
> the hard disk
> versions IIRC).

I could be wrong about the LSI-11...perhaps it was an 11/23. I don't
know my 11 stuff very well...something seems to recall a "J-11" chipset
that made up the CPU? Was that the 11/23, or LSI 11?

>
> > have multiple sessions running. There was an emulator subsystem
> > that could be connected into the system to allow emulation under
> > control of the TNIX environment. As I remember, the emulator was
> > another unit about the size of the 8650 that connected in through
> > a special interface, that provided an in-circuit emulation
> capability
>
> Mine has a few cards in the second box (the one not
> containing the PDP11
> or drives) that are an In-Circuit Emulator for the TMS9905.
>
Yup...that's exactly how I remember...
>
Re: Micropolis Drive interface
> Yes. I have a reasonable amount of data on these drives. The
> bare drive
> (without the controller) was used in the PERQ 2T1, which is
> where I first
> came across it.
>
Really? Didn't know that these drives were used in PERQ machines.
I remember when PERQ came out with their first machine. A big
demo was put on for the Tektronix Scientific Computer Center staff (of
which I was a part).

PERQ came in and showed their machine. I remember thinking
how it seemed like the old Tek Magnolia (which predated the PERQ
by a few years, as I recall) seemed like it was as good as the PERQ.
I always thought Tek was really dumb for now following on with Magnolia.
Anyway, the thing I thought was so cool about the PERQ was the
writable control store. Pretty amazing.

> The bare drive has a 50 pin interface connector. Basically,
> the interface
> is 8 data lines, 2 address lines, strobe, etc to load
> cylinder and head
> numbers, control various drive functions, etc. And a seperate 'raw'
> datastream for the actual data.
>
I didn't have any data about the bard drive interface, all the data
I had was for the Micropolis formatter board.

> The controller board sits on top of the drive and gives you a nice
> formatted interface to the host. It converts between the raw
> data from
> the drive and nice 8 bit bytes for the host. I can't remember
> much about
> that,

As I remember, there was a special jumper block on the formatter board
that selected the number of bytes/sector...and I needed 256 bytes/sector
for my OS-9 driver, and the drives were shipped with the jumper block
configured for 512 bytes/sector. I did some serious
reverse-engineering to figure out how to change the jumper block
to allow for 256 byte/sector operation. I remember feeling quite proud
of myself when I wired up a new jumper block and tried it, and low
and behold, it worked!

Rick
Received on Mon Nov 22 1999 - 19:25:22 GMT