PDP-11/03, LSI-11 KEV11-C CIS option

[Eric Smith]

On Sun, Jan 31, 2016 at 11:38 AM, John Wilson <wilson at dbit.com> wrote:
> On Sat, Jan 30, 2016 at 05:16:18PM -0700, Eric Smith wrote:
>>Just looking for the CIS chip(s), and just to borrow long enough to
>>dump the microcode.
>
> This is the first I'm hearing that there was a CIS micROM option for
> the LSI-11[/2].  I'm very impressed if they were able to pull that off!

I think the KEV11-C is not full CIS, but rather a subset. It's usually
described as DIBOL Instruction Set (DIS). There are two diagnostics
for DIS; VKAI?? is for move and string instructions (MOVC, MOVRC,
CMPC, LOCC, SKP, SCANC, SPANC), and VKAJ?? is for decimal instructions
(ADDN, SUBN, CMPN, CVTNL).

DIS was standard in some of the D3xx commercial-oriented LSI-11
systems, such as the D324. The KD11-P and KD11-Q processors are the
M7264-BB and M7264-YB, respectively, with the KEV11-C DIS
preinstalled.

It's not possible to use DIS together with the KEV11-A EIS/FIS (or
KEV11-B EIS) options, due both to limited number of MICROM sockets,
and that the base instruction set and DIS together take up the entire
microcode address space. However, the LSI-11 Systems Service Manual
Volume 3 (Jan. 85) states that the VKAB?? EIS diagnostic will work
with the KEV11-C, so presumably KEV11-C includes both DIS and EIS (but
not FIS).

My understanding is that the KEV11-C option consists of two non-hybrid
MICROMs, 23-004B5 and 23-005B5, and is used with a 40-pin hybrid
containing the two MICROMs of the base instruction set. The hybrid is
23-001B6, 23-002B6, or 23-003B6; I don't know what differences there
are between the three hybrid part numbers, but they're described as
being for M7264 ECO 10, ECO 12, and ECO 16, respectively.

> I have the QIL CIS micROM for the LSI-11/23, and I too have been meaning
> to build a ROM dumper for it, since it's a rare part but could be easily
> emulated with modern hardware on a daughterboard (easier than the 11/44
> anyway, where the CIS option means extra boards).

I've got bad news for you. The control chips of the F11 chipset aren't
just ROMs; they include the control sequencer (one per control chip).
Unless there's an undocumented test mode, it's not possible to dump
the full microcode word, since only part of it is output to pins, and
it's not possible to dump the contents purely sequentially. I think it
will be necessary to dump them optically, which of course requires the
chip carriers to be delidded. While this is not necessarily
destructive, it can't be 100% guaranteed not to damage the chips, and
even if it doesn't, the future reliability of them may be diminished.

On the other hand, as you say, if we could dump the KEF11-B CIS
control chips (optically or otherwise), we could build a replacement
daughtercard that emulates KEF11-B, and then anyone that wanted to
have CIS on their 11/23 or 11/24 would have a relatively low-cost
option to do so.

I'd really like to get the F11 base instruction set control (part of
the DCF11 hybrid) and the KEF11-B CIS (six chip hybrid spanning two
40-pin DIP sockets) to be dumped in this way, but only if someone is
willing to sacrifice them. Given that the KEF11-B is fairly uncommon,
I'm not holding my breath...