Jay West wrote:
TSIA :)
I was skimming through the 11/34 manuals, and some confusion about the
termination of that system made me re-read through the M9312 manual.
Something doesn't seem right and points to me not understanding
something. I seek enlightenment from the list.
As I've understood things, you have to terminate both ends of the
unibus. Let's say you're dealing with an 11/34 with no expansion
backplanes (ie. just a DD11-PK). According to the 11/34 manual you put
the M9302 terminator in slot 9AB. That makes sense, it's the unibus
out. All the other AB slots are MUD (except slot 1AB, and we'll get to
THAT in a moment). So about that other end of the bus now! The manual
states you need a M9301 bootstrap/terminator just after the processor
(typically, it may be further down if you have the FP and/or Cache
options). That also makes sense to me, gotta terminate that "near" bus
end. I had a M9302, but I didn't have a M9301. I wondered if I could
use a M9312 in place of the M9301. So...
I skimmed the M9312 bootstrap/terminator manual, and sure enough, it
definitely talks about being able to use the M9312 instead of a M9301
in a processor that "has bus termination built into the cpu such as
the 11/04 & 11/34". It even says that you must set the jumpers on the
M9312 to remove it's termination capability, because in these specific
processors the cpu provides bus termination. Ok, great. I can use my
M9312 in place of the M9301 long as I jumper it as described. But...
now wait a second! This gives rise to my first question. If the cpu is
providing termination, and if you are going to use a M9312 on the near
side you have to "jumper off" the M9312's termination ability... then
WHY is a M9301 called a "bootstrap TERMINATOR". It must NOT be a
terminator, just a bootstrap rom card. Right? It seems to me that if
the cpu is providing the termination on the near end, that it is
always doing so. Not JUST when you pull out the 9301 and put in the
9312. What gives there?
That gives rise to my second question. IF the cpu is providing unibus
termination on the one side... why can't you jumper the M9312 to
provide bus termination and just put it in slot 9AB? You would get the
other end (unibus out) terminated AND get your boot roms as well. This
way the M9312 would replace both the 9301 AND the 9302. A few people I
mentioned this to said you can't do that. But given the above... it
seems logical. What's the scoop here?
Clarification is much appreciated!
Jay West
The issue is there are two different aspects of the UNIBUS that are
being terminated:
The 11/04 and 11/34 CPUs *ONLY* electrically terminate the BG1-4 and NPG
lines; the rest
of the UNIBUS address/data/control lines are unterminated and have
drivers/receivers only.
The M9301/M9312 modules *ALWAYS* electrically terminate the UNIBUS
address/data/control
lines, *EXCEPT* there are jumpers (W1-W5) that allow selection of
whether or not the
UNIBUS BG1-4 and NPG signals are electrically terminated.
The M9302 module *ALWAYS* terminates ALL UNIBUS lines,
address/data/control, BGx/NPG.
It also does SACK turnaround, but that is a hole other issue.
So when the M9312/M9301 manuals talk about 'programmable termination'
they really only
mean 'BGx and NPG'; the termination of the other UNIBUS signals is not
programmable.
What really happens is:
Slot1/M8265 CPU - unterminated UNIBUS addr/data/ctrl
Slot2/M8266 CPU - terminated UNIBUS BGx/NPG drivers (pullup only)
Slot3/M9312 BOOT/TERM - terminated UNIBUS addr/data/ctrl; BGx/NPG
unterminated/passthru
....
Slot9/M9302 TERM - terminated UNIBUS addr/data/ctrl; terminated BGx/NPG
So electrically the UNIBUS is terminated by the M9312 in slot 3 and the
M9302 in slot 9.
For the specific case of the BGx/NPG signals, they are terminated
(pullup only) at the
processor in slot2 (or slot1 for the 11/04) and slot9 at the end of the bus.
Hope this clears things up a bit.
Don North