No gaps between quad modules?

[Bob Armstrong]

I was reading

http://www.bitsavers.org/pdf/dec/pdp8/omnibus/EK-0PDP8-SP-001_PDP-8_Family_Configuration_Guide_Apr78.pdf

and it has some interesting configuration rules for 8/A systems. In particular, in an 8/A system the hex modules are supposed to go in the top of the backplane and the quad options at the bottom - I assume that's because only the top slots have hex wide (or really "penta wide") connectors; the bottom ones are all quad.

However it also says "there can be empty slots between hex modules but there should be no empty slots between quad modules..." That seemed odd - the OMNIBUS doesn't have any daisy chained signals, so why do empty slots matter? And why only between the quad (8/E style) modules? Did the 8/E have the same rule?

[Robert]

On Thu, 18 Jun 2020 17:20:30 -0700 (PDT), Bob Armstrong <b...@jfcl.com>
wrote:

No idea why it would say that. No the 8/e never had such a 'rule'
AFAIK and it's pretty common to have gaps all over the place IME
especially if there are a few peripherals with lots of cabling.
FWIW my 8/a seems happy enough with gaps between the disk set and the
papertape interface cards. Not heard of this idea before.

[Don North]

The OMNIBUS (8/efm) has certain ordering constraints on module placement:
(1) console goes in slot 1 (obviously)
(2) CPU timing module goes in slot 2
(3) bus loads module goes in the last slot in the backplane
(4) core memory array(s) go in slots just adjacent to bus loads module
(5) RFI shield goes in slot prior to first core memory array
(6) CPU modules go in slots in front adjacent to the timing module
(mainly because all the over the top connectors force this to happen)
(7) peripheral placement is generally unrestricted, spaces between modules ok

I have a PDP-8m and the 20 slot backplane is basically full, maybe one open
slot. CPU w/EAE and two core stacks. Floppy and TTY interface.