VAX 11/750 rescued, alas...

Ethan Dicks ethan.dicks at gmail.com
Sat Oct 6 03:28:00 CDT 2007


On 10/6/07, J Blaser <oldcpu2 at rogerwilco.org> wrote:
> The gods smiled upon me, and I was chosen to rescue the DEC
> VAX 11/750 that Richard alerted us too about a month ago...
>
> http://www.rogerwilco.org/VAX11-750
>
> ... To my horror (it is October, after all), I find just
> a single board installed in the CPU backplane, and one loose
> q-bus board!

That is unfortunate.

> The first is a System Industries 9700-6301
> with the following significant chips: Signetics N8X60N, AMD
> AM9128 (x2), Motorola MCM93L422PC (x4), TI SN74S181N, and
> about a dozen TI 82S137s with little numbered stickers on
> each.  Otherwise it's loaded with TTL logic chips.  There
> are no headers or other connectors on board, just the
> backplane fingers.   I have no idea what this board is.  I
> openly admit that I'm a complete {non-uVAX | massbus |
> unibus} novice.

That looks to me (from appearance and the part number) to be the
11/750-specific SI-9900 interface.  I have one in an 11/750 in my
quonset hut.  SI made a number of host-specific interfaces (Qbus and
11/750, for certain, others likely)... this one fits in the same kind
of slot as a DEC RH-750 MASSBUS interface, but is not, itself,
specifically a MASSBUS card (there is at least one other board for
that slot include, an additional UNIBUS interface).

For the case of what you have there, you should also have a pair of
40-pin cables going off the backplane for that slot, out to an I/O
bulkhead header.  There would then be a pair of (possibly shielded)
40-pin cables running off to a 5.25"-tall rackmount box marked "SI
9900".  It should have at least one host board to receive the pair of
40-pin cables, then on the other side of the box, one or more disk
interfaces (there should be room for two or three of those).  We had a
pair of SMD interfaces in ours, attached to 160MB and 450MB Fuji
disks, emulating a pair of RM03s and a "custom-sized" RM05,
respectively.  The 450MB disk always caused us heartburn because we
had to patch the, IIRC, DBDRIVER.EXE to change the RM05 geometry,
every time we upgraded VMS.  The 160MB drive emulated a pair of
RM03s/9762s, with no tweaking.

I did notice you have a cable hanging down for a UDA50... it's likely
that they had both SI disk _and_ Unibus-attachd SMD disks.  That was a
common arrangement for a variety of reasons.  You will also want to be
on the lookout for a DU boot ROM.  The 11/750 didn't use the TU58 for
boot files as the 11/730 and 11/725 did.  You just twisted the ROM
selector to the right device, then pressed reset/boot.  the D ROM was
typically the TU58 boot (DDA0?) so you could boot tape-based
diagnostics.

> The second board is an MCD MLSI PC-11,
> which I'm guessing may be some kind of dual parallel
> interface.  I'll have to do a little scouting on Bitsavers
> and Manx to see if I can turn up anything useful.

If you don't find anything, the board looks to me like a papertape
punch/reader interface.  Try throwing that into your searches.

> Anyway, sadly, this box is not much more than an empty
> carcass, without any brains.

Good luck on finding boards.  They are out there, but I doubt they'd
be inexpensive from a dealer.

If you do find boards, you will have to match up the memory to your
memory controller.  There are three flavors.  The oldest only can talk
to 256KB boards (the same ones that work in an 11/730 or 11/70).  You
can stuff a max of 2MB in there - enough for ancient UNIX or VMS, but
I wouldn't want to run VMS 4.0 or newer on that.  The most common
version will support eight 1MB boards (same as used with an 11/730 or
11/725).  The newest version supports a pair of 4MB boards and a six
1MB boards (14MB).  I have never worked with that one, but it would be
something to look for.

The other thing to do, and I don't have the docs handy for this, is to
check your backplane in the memory area, to confirm what it's wired to
accept.  I have 11/750 S/N BT0000354.  It shipped with 512KB, then I
upgraded it to 8MB by adding the extra refresh/mux line and changing
the SID dipswitch (rather than paying DEC to do it).  I presume an
additional wire is needed for the 14MB mod.  I think you should be
able to put a lower-rev memory controller in a backplane designed for
a newer one, but you might need to update the SID switch to match what
the OS is expecting.

Don't let this worry you until you get enough boards to throw together
into a CPU... just get some memory - 1MB boards _should_ be safe -
then worry about which memory controller you find next.  256KB boards
shouldn't be expensive, I wouldn't think - since nobody wanted them
much after 1990.  Everyone wanted 1MB boards for several varieties of
VAXen, so lots of those were produced, too.  We even had a couple
3rd-party boards with a handful of 256Kbit DRAMs, but the DEC boards
(packed solid with 4164s) were the easiest to locate at surplus
dealers.

> Now, I should
> say that the donor was extremely nice, and offered to pass
> along anything related that eventually turns up as they dig
> deeper into their massive pile (12' x 15' x 20', literally
> boxes/cartons/PCs upon boxes/cartons/PCs), but I'm not sure
> I should hope for much.

Have them keep an eye out for that SI disk box - it's the other half
of the board you already have.

> Am I foolish to ask?  Anyone with a spare set of VAX 11/750
> modules?

I have no spares, sorry.  Just a couple of machines and lots of
experience crawling around inside them (at Software Results, I was our
own primary hardware maintenance resource from about 1990-1994).

-ethan

P.S. - don't forget to pick up a short length of 1/2 ID Tygon tubing
to rehab the TU58 roller.  If you _do_ eventually have to scrap out
the machine, save the backplanes, PSUs and TU58 for other 11/750
owners.  Best of luck on finding a boardset, though.  The 11/750 is a
really neat beast.  It would be good to see you getting it going.



More information about the cctech mailing list