23 Oct
2021
23 Oct
'21
7:16 p.m.
Anyone know about the Dec DMX11? It was apparently a 64 serial line Mux
that plugged into pdp11 Unibus systems and had a fair amount of both
intelligence and insanity on board. Reason is I'm looking into some old
documentation about late 1970's Tote systems for Asian race tracks and
the system they were building was beyond astronomical in terms of insanity.
(As in an 8 way pdp11/70 system cluster, with 11/04's running these
DMX11's to hundreds of terminals)
25 Oct
25 Oct
11:16 a.m.
On Oct 23, 2021, at 10:16 PM, Chris Zach via cctalk
<cctalk at classiccmp.org> wrote:
Anyone know about the Dec DMX11? It was apparently a 64 serial line Mux that plugged into
pdp11 Unibus systems and had a fair amount of both intelligence and insanity on board.
Reason is I'm looking into some old documentation about late 1970's Tote systems
for Asian race tracks and the system they were building was beyond astronomical in terms
of insanity.
(As in an 8 way pdp11/70 system cluster, with 11/04's running these DMX11's to
hundreds of terminals)
The April 1983 Options & Modules List (see
http://www.bitsavers.org/pdf/dec/modules/modulesAndOptions/) shows it on page 156. It is
listed as a CSS product, 64 line mux controller (DMA both ways, so similar to the DH11)
with separate 64-port line cards. Three flavors of line card: 20 mA loop, "EIA"
(presumably that means RS-232) and "Differential" -- perhaps RS-422?
Date is given as 08/1980, but that's the last update of the entry. Status is
"6". The April 1983 document is missing its cover material which explains the
codes, but the December 1975 edition (same Bitsavers directory) lists them. "6"
means "Obsolete, but can still be custom-built".
Given that it's a CSS device, documentation for it is likely to be hard to find, and
software for it even harder. For example, while RSTS would be a natural OS to support
this (given that it can handle 128 terminal lines) it doesn't and as far as I know
never has.
paul
12:22 p.m.
Thanks! I found the PCL11 in the 1980 Terminals and Communications
handbook, but the biggest thing they had was the COMM/IO/DUP
micro-controller that ran up to 6 DZ11's for 48 terminals. And that was
in the Customer Specific Solutions section.
No mention of what the DMX11 was, but it must have been a beast: One
thing I do see in this report is that DEC went nuts because the customer
was holding them to an interesting set of requirements that changed
mid-project.
From what I can see, it was some sort of a "block mode" device that
could support 64 terminals at a shot. Then the client/DEC also wanted it
to drive VT52's in character mode as well because the "Terminal
Concentrator Device" (no idea what that was) was not going to be
supported by RSX11/D going forward. It was still supported in RSX11/M
now I'm wondering what THAT device was as well.
Anyway trying to stuff character mode and block mode in the same
controller doubled the microcode requirements. Worse, they couldn't lay
out the components on a nine board layout and the customer refused to
allow more boards to make it fit. So it had "problems".
So what was a TCD? Was that a pdt11/150 by chance?
C
On 10/25/2021 2:16 PM, Paul Koning wrote:
On Oct 23, 2021, at 10:16 PM, Chris Zach via
cctalk <cctalk at classiccmp.org> wrote:
Anyone know about the Dec DMX11? It was apparently a 64 serial line Mux that plugged into
pdp11 Unibus systems and had a fair amount of both intelligence and insanity on board.
Reason is I'm looking into some old documentation about late 1970's Tote systems
for Asian race tracks and the system they were building was beyond astronomical in terms
of insanity.
(As in an 8 way pdp11/70 system cluster, with 11/04's running these DMX11's to
hundreds of terminals)
The April 1983 Options & Modules List (see
http://www.bitsavers.org/pdf/dec/modules/modulesAndOptions/) shows it on page 156. It is
listed as a CSS product, 64 line mux controller (DMA both ways, so similar to the DH11)
with separate 64-port line cards. Three flavors of line card: 20 mA loop, "EIA"
(presumably that means RS-232) and "Differential" -- perhaps RS-422?
Date is given as 08/1980, but that's the last update of the entry. Status is
"6". The April 1983 document is missing its cover material which explains the
codes, but the December 1975 edition (same Bitsavers directory) lists them. "6"
means "Obsolete, but can still be custom-built".
Given that it's a CSS device, documentation for it is likely to be hard to find, and
software for it even harder. For example, while RSTS would be a natural OS to support
this (given that it can handle 128 terminal lines) it doesn't and as far as I know
never has.
paul
937
days inactive
938
days old
2 comments
2 participants
participants (2)
-
cz＠alembic.crystel.com -
paulkoning＠comcast.net