21 Oct
2020
21 Oct
'20
9:36 a.m.
From: Jay Jaeger
> 2--M7134 KT24 Memory map
M7891 UNIBUS Memory (256K, I think, presumably
addressed for 0).
If that's all the memory you have, the KT24 isn't really doing anything
(well, monitoring power; holding boot PROM's; etc). Is your MS11-L configured
to be EUB memory, then? If no KT24 is plugged in, the CPU detects that there
isn't one there, and permanently, statically maps the UNIBUS straight across
to the bottom 256KB of EUB space. (Presumably a low-cost option for the /24.)
Although the 'straight across' mapping only applies to UNIBUS->EUB cycles,
not EUB->UNIBUS cycles, such as those from the CPU; but the TM says [pg 2-31
in the 003 version] that if E124-S6 is OFF, "the lower 18 bits of every
address go to the UNIBUS", which implies that when OFF, UNIBUS memory appears
at 0 in the CPU's address space. So it should work as UNIBUS memory, with
E124-S6 OFF; it would be interesting to verify that. The TM also says (pg.
2-40) "systems with UNIBUS memory ... require changes to be made to the
mapping jumpers [on the KT24]".
Let me know if you want me to go thru the process of
..
2) Pull the boards and document
What kind of box is your -11/24 in, a BA11-A, or BA11-L?
If the latter, I'd be really grateful for some closeups of the interior, so I
can put mine back together, and do some of these experiments. (Yes, yes, I
know I should have taken pictures before I took it apart; I was just
starting, and was going by how we used to do things, back before there were
digital cameras.) I could probably work it out by staring hard, and thinking
harder, along with the prints, but photos would be a lot easier! :-)
If it's a BA11-A, I'm still trying to get an image of the special power
adapter used to turn the bus bar of the BA11-A into the 6/15 pin Mate-N-Lok
connectors used by the -11/24 backplane.
Noel
21 Oct
21 Oct
5:28 p.m.
If that's all the memory you have, the KT24
isn't really doing anything
(well, monitoring power; holding boot PROM's; etc). Is your MS11-L configured
to be EUB memory, then? If no KT24 is plugged in, the CPU detects that there
isn't one there, and permanently, statically maps the UNIBUS straight across
to the bottom 256KB of EUB space. (Presumably a low-cost option for the /24.)
Yes. I re-jumpered the memory to be EUB memory (bit of a pain as I
wicked the solder, then moved the jumper) I could move it back but I
think the problem is before this (and remove the Unibus map) however in
that case do the five slots under the CPU board go from being EUB to MUD
slots? Or are they blanked out with H727 knucklebusters?
What kind of box is your -11/24 in, a BA11-A, or
BA11-L?
The little 5.25 inch one. I just put up a few pics at
www.crystel.com/pdp, what kind of shots are you looking for?
CZ
22 Oct
22 Oct
8:02 a.m.
On 10/21/2020 11:36 AM, Noel Chiappa via cctalk wrote:
No idea how it is configured as it has always "just worked". Factory as
a guess. Will pull it and document it.
From: Jay
Jaeger
> 2--M7134 KT24 Memory map
M7891 UNIBUS Memory (256K, I think, presumably
addressed for 0).
If that's all the memory you have, the KT24 isn't really doing anything
(well, monitoring power; holding boot PROM's; etc). Is your MS11-L configured
to be EUB memory, then? If no KT24 is plugged in, the CPU detects that there
isn't one there, and permanently, statically maps the UNIBUS straight across
to the bottom 256KB of EUB space. (Presumably a low-cost option for the /24.)
Although the 'straight across' mapping only
applies to UNIBUS->EUB cycles,
not EUB->UNIBUS cycles, such as those from the CPU; but the TM says [pg 2-31
in the 003 version] that if E124-S6 is OFF, "the lower 18 bits of every
address go to the UNIBUS", which implies that when OFF, UNIBUS memory appears
at 0 in the CPU's address space. So it should work as UNIBUS memory, with
E124-S6 OFF; it would be interesting to verify that. The TM also says (pg.
2-40) "systems with UNIBUS memory ... require changes to be made to the
mapping jumpers [on the KT24]".
It's the BA11-A - 10.5" flavor. I'll take photos you requested here and
document the board jumpers and switches over the next week or so, for my
own use, if nothing else.
JRJ
Let me know if you want me to go thru the process
of
..
2) Pull the boards and document
What kind of box is your -11/24 in, a BA11-A, or BA11-L?
If the latter, I'd be really grateful for some closeups of the interior, so I
can put mine back together, and do some of these experiments. (Yes, yes, I
know I should have taken pictures before I took it apart; I was just
starting, and was going by how we used to do things, back before there were
digital cameras.) I could probably work it out by staring hard, and thinking
harder, along with the prints, but photos would be a lot easier! :-)
If it's a BA11-A, I'm still trying to get an image of the special power
adapter used to turn the bus bar of the BA11-A into the 6/15 pin Mate-N-Lok
connectors used by the -11/24 backplane.
Noel
13 Nov
13 Nov
3:16 p.m.
New subject: 11/24 board settings (Re: Next project: 11/24. Does it need memory?)
On 10/21/2020 11:36 AM, Noel Chiappa via cctalk wrote:
CPU E124 S6 is indeed ON, so if the TM is right, only the high addresses
go to the UNIBUS. Skeptical of that, unless by UNIBUS they mean "UNIBUS
other than memory".
UNIBUS MAP no "TP" wires, so the lowest UNMAPPED address would be
740 000 if the TM is accurate. The only wire in is W13 (ROM enabled)
near as I could tell.
Labeled M7891-BB. Presumably 128KB (64KW).
The M7891 (MS11-L) is addressed at 0 as near as I can tell - S2 (E61)
switches are all ON, and it appears to be set for the default 4K I/O
Page from 124K to 128K.
My manual EK-11024-TM-001 (not -003).
Yes, it is a BA11-A (10.5" tall). Do you know offhand where that power
adapter is? I fear you'd be asking me to disassemble the backside of
the power supply, or, worse, take the power supply out, which I am not
inclined to make happen. ;)
There is a schematic of what I think is the board you are referring to
on bitsavers at
http://bitsavers.org/pdf/dec/pdp11/1124/MP01018_Rev_F_11-24_Engineering_Dra…
PDF page 161 of 200 (5413815)
From: Jay
Jaeger
> 2--M7134 KT24 Memory map
M7891 UNIBUS Memory (256K, I think, presumably
addressed for 0).
If that's all the memory you have, the KT24 isn't really doing anything
(well, monitoring power; holding boot PROM's; etc). Is your MS11-L configured
to be EUB memory, then? If no KT24 is plugged in, the CPU detects that there
isn't one there, and permanently, statically maps the UNIBUS straight across
to the bottom 256KB of EUB space. (Presumably a low-cost option for the /24.)
Although the 'straight across' mapping only applies to UNIBUS->EUB cycles,
not EUB->UNIBUS cycles, such as those from the CPU; but the TM says [pg 2-31
in the 003 version] that if E124-S6 is OFF, "the lower 18 bits of every
address go to the UNIBUS", which implies that when OFF, UNIBUS memory appears
at 0 in the CPU's address space. So it should work as UNIBUS memory, with
E124-S6 OFF; it would be interesting to verify that. The TM also says (pg.
2-40) "systems with UNIBUS memory ... require changes to be made to the
mapping jumpers [on the KT24]". Let me know if you want me to go thru the process
of
..
2) Pull the boards and document
What kind of box is your -11/24 in, a BA11-A, or BA11-L?
If the latter, I'd be really grateful for some closeups of the interior, so I
can put mine back together, and do some of these experiments. (Yes, yes, I
know I should have taken pictures before I took it apart; I was just
starting, and was going by how we used to do things, back before there were
digital cameras.) I could probably work it out by staring hard, and thinking
harder, along with the prints, but photos would be a lot easier! :-)
If it's a BA11-A, I'm still trying to get an image of the special power
adapter used to turn the bus bar of the BA11-A into the 6/15 pin Mate-N-Lok
connectors used by the -11/24 backplane.
Noel
1499
days inactive
1522
days old
3 comments
3 participants
participants (3)
-
cube1＠charter.net -
cz＠alembic.crystel.com -
jnc＠mercury.lcs.mit.edu