On Jan 4, 13:35, Ethan Dicks wrote:
> --- Tom Leffingwell <tom_at_sba.miami.edu> wrote:
> >
> > I have a PDP 11/23 (M8186 CPU with floating point and MMU options)
> > in a 4 slot BA11-MA box that at some point in its life was an 11/03.
It
> > had an M8044-DF 32k memory module, which I'm trying to replace with a
> > 128k M8059-KJ. The system works fine with the 32k module, but won't do
> > anything with the 128k module.
>
> Have you checked the jumpers on the M8059? Where in memory does it think
> it is supposed to live? Even the M8044 has address jumpers.
Two other things occur to me. The first is that memories like the MSV11-D
(M8044) and MSV11-L (M8059) have jumpers to enable or disable memory that
corresponds to the I/O page. This gets a little complicated... Most DEC
memories decode the BBS7 signal to sense access to the I/O page and disable
memory access accordingly. The MSV11-D has a jumper 1-2 to enable memory
in the lower 2KW of the I/O page for systems that don't use the whole I/O
space; jumper 2-3 (the factory setting) to disable memory in the whole top
4KW. The MSV11-L has a similar arangement, using jumpers 28-29 to enable
the memory in the bottom half of the I/O page, and 27-28 to disable it.
> > I'm not familiar with PDP-11's, but it seems like my backplane is only
> > 18-bit, while the new memory module is 22-bit.
>
> Yes, I would expect that your backplane is 18-bit. It wouldn't matter,
> anyway - 18 bits is 256Kbytes (2^18 = 262,144) or 128Kwords. The 128K
> card will fill your memory space, but it should work on an 18-bit
> blackplane. You can also stuff 4 M8044 cards in there - 32Kwords each
> for a total of 128Kwords.
And that's the second thing. The MSV11-L doesn't use BBS7 for everything
it decodes, so you have to set it according to whether it's in a 22-bit
system or an 18-bit system. If there's a jumper from R-T, it's set for a
2MW system. Remove it for 128KW systems.
It's also a parity memory, unlike the MSV11-D, and furthermore it has it's
own on-board parity control and reporting register. I don't think that
would stop it working if not set up, though. If you want to check the
parity jumper settings, the factory default is 9-10 out, 10-11 in, 18-19
out and 19-20 in (to enable parity); 1-2 in and 2-3 out (to enable register
reporting); 6-7 in and 7-8 out (necessary to disable wrong-parity writing
which is used for testing); F-H in and J-H out (to enable the CSR for
parity setting/reporting rather than other methods).
There are also jumpers to enable/disable half the board. Normal settings
are 32-33 out and 33-34 in for a fully-populated board (reverse for
half-populated) and 15-16 in and 16-17 out (the reverse disables the lower
bank, used if part of the bank fails).
Most of the remaining jumpers set the starting address. For a starting
address of zero, you want pins P,N,M,L disconnected, and pins V,W,X,Y,Z
disconnected (the addressing is done in two parts). To change the address
to something other than zero, you would connect one or more of these pins
to pin K (ground).
> > I've also read that the M8186 board is only 22-bit compatible after
> > revision C. I can't find any mark on the board showing what revision
it
> > is. Is there another way to tell?
>
> Is it on the maroon handle? I don't think it's anything in the solder
> mask/copper; there might be an ink stamp with the revision or perhaps a
> sticker. If you can't tell in anyway, shape or form, perhaps you have
> a rev A.
Quick check is to hold the board with handles at the top/contacts at the
bottom, and look for a metal jumper (W18) about an inch below the handles,
about halfway between them, and above a diode jumpered with a red wire. If
it's there, it's a Rev.A.
> > Also, is possible to modify the 18-bit bus and make it 22-bit, or
> > maybe by swapping out the backplane?
>
> You can run the extra backplane wires. I have done it. Having done so,
> you may have to find a way to terminate them. Newer boxes are already
Q22
> and have bus termination built-in. Older backplanes depended on a
> termination card - the BDV-11 is one that is termination plus bootstrap
> ROMs.
I've done it too. A standard BDV11 doesn't terminate the upper 4 address
lines, but there's an ECO (ECO 005?) to fix that. It consists of adding 4
short wires from the contact fingers to termination resistors that are
otherwise unused.
--
Pete Peter Turnbull
Network Manager
University of York
Received on Fri Jan 04 2002 - 19:05:02 GMT