24 Sep
2019
24 Sep
'19
8:52 a.m.
On 9/24/2019 3:15 AM, Tony Duell via cctalk wrote:
On Tue, Sep 24, 2019 at 9:13 AM Bill Degnan via
cctalk
<cctalk at classiccmp.org> wrote:
You confirmed it is not Omnibus bd, like a drive bootstrap board?
I think with that many address switches, a 'vector' switch, the grant
jumpers and the connector at that end it's Unibus.
-tony
Assuming UNIBUS...
Looking at the address switches, I see 11 selectable bits out of how
many we don't know.
If at one extreme, the bits start at bit 15 (MSB) [assuming bits 16 and
17 are also 1], then the address of the ports on the board would be
77744X - 77747X - that seems like a lot. (They might start at bit 17,
but then the space used by the board gets absolutely crazy).
If, on the other extreme, we suppose that the top 4 bits are 17 (along
bit bits 17 and 16), then we get 777762 - 777763. That might actually
be possible.
One might continue the exercise one bit at a time between the two to see
if the address lands anywhere "interesting".
777762
777744
777710 Not here (Conflict with CPU registers)
777620 Unlikely (memory management)
777440
A more sure way would be to find the address line to which the first
address switch corresponds.
JRJ
Here's the full board, for reference:
http://yahozna.dyndns.org/scratch/random/board1.jpeg
Thanks,
Josh
/Mattis
