On 22/05/2007 20:23, Johnny Billquist wrote:
"Rod Smallwood" <RodSmallwood at
mail.ediconsulting.co.uk> skrev:
Well
that's seems to define the 11/84.
How does this all apply to the 11/94
(which is the real problem)
Well, the UBA is the same, so the same PMI Unibus signals are originated
at that end.
No memory can exist in an 11/94 though, since all the memory are on
board. However, the communication between the CPU and UBA is using the
PMI protocol, still.
However, the dealings with interrupts and so on must still be done in
the same way as for the 11/84 as well.
So I would say that the 11/94 behaves exactly like the 11/84.
Even the backplane is the same, at least the same as the 11/84-E and
11/84-D systems.
(Do anyone know if an 11/94 with 2 megs on board could
use an external 2
megs?)
Since all the relevant QBus and PMI signals must be on the CPU card, in
order that the KDJ11 and the KTJ11 can communicate, and everything else
is the same as an 11/84, it ought to work. The only issue I can think
of is that the 11/94 uses parity memory, whereas the most common 11/84
memory cards are MSV11-J ECC memory. Presumably using MSV11-RA parity
memory would get round that, if it were a problem.
--
Pete Peter Turnbull
Network Manager
University of York