(I've got to get things going so I can login at night and read/respond
to everything sooner! We may have the solution.. at bottom )

No, I've definately had the rd54 and an rd53 in the BA23 at the same time.
The only reason I don't now is that I'm trying to spare the 53 further
grief, and hopefully I'll be able to pull data off it at a later time.

As Allison pointed out, drive 1 addressing should be 3 on the 0 to 3 scale,
and drive 0 should be addressed at 2. And of course, the RX50 module has
to be yanked, hanging anything outside the chassis is a no-no. I've been
playing with the 2 empty drive bays, cables pulled from the clearly marked
'Fixed Disk 0' and 'Fixed Disk 1' connectors.

The RD31's have been formatted on an RQDX3 installed in a PDP-11 using
XXDP and ZRQCF0 (I know it well enough to type in my sleep). For each
I *told* it this was an RD31, since the autoformat didn't like it and
couldn't find the bad block map. Both NetBSD and VMS like the drive
when it's in an external drive chassis, and attached via the RQDX3
*AND* an RQDXE, which is the only way to expand. By the way, in the
external enclosure the address on the drive is 3, making it internal
at 3, and external at 3. This config always boots, always works correctly.
The long term goal is to get 3 drives going, meaning the short term
goal is to get the 2 internal ones going. (Technically, the external
chasis handles 2 drives, but I'd like to discuss that expansion & cableing
in another thread.)

I suspected the terminator was hard-soldered on the non RD31 drives.
I traced the board some and found the resistor pack right there.

I know I've had the RD53 or RD54 as my boot drive at address 2,
and I'm sure at some point I've put the RD53, RD52, RD51 in as address 3
for drive 1. It's these 31's which simply do not want to play nice.


Dave Clarke may be on to something, Megan too. I've got the 4 switch
panel on both the vVax and pdp11. Odd though that I am able to get
the RD5n drives to cooperate. I'll review the archives to see if
anybody ever commented on how to re-engineer the extra logic.

I'll follow that route and let everyone know what the results are.

Zach