PDP 11/73 in BA213 Cabinet Error

[Hunta]

Thus it really is an 11/83 CPU

CPU is an J-11, CPU board is M8190 (KDJ11-B) and the system as a whole is PDP-11/83 or PDP-11/84

The first four slots in an 11/84 are Qbus slots.

The first three slots are QBus slots and fourth slot - united QBus-Unibus for KTJ11-B.

 

[Hunta]

To be more precise, the first three slots have the left half (AB) is QBus, the right half (CD) is PMI, the fourth slot AB part is QBus, and EF part is Unibus.

 

[tradde]

Wasn't sure about the 4th slot. Should have been obvious it had to be different to be able to go from Qbus to Unibus, and back. I don't remember if it reports itself as an 11/83 or 11/84.

 

[Hunta]

 

[Hunta]

Slot 0 - MDM
Slot 1 - M8190
Slot 2 - M8637 (PMI RAM)
Slot 3 - M7556 (minimum load module)
Slot 4 - M8191 (KTJ11-B - qbus-unibus converter)

 

[gslick]

Formally, the board type is identified by the contents of the maintenance register (17777750) - code 0100 - PDP-11/53, 0001 - PDP-11/73, PDP-11/83(84) - 0010, PDP-11/93(94) - 0101

The Maintenance Register in the M8192 KDJ11-A, M8190 KDJ11-B, and M8981 KDJ11-E also has a bit (Bit 8) indicating whether or not the Floating-Point Accelerator is installed, and in the M8190 KDJ11-B and M8981 KDJ11-E a bit (Bit 9) indicating whether or not the board is installed in a UNIBUS system.

If I follow the schematics in the maintenance print sets correctly, the UNIBUS System bit in the M8190 KDJ11-B and M8981 KDJ11-E is logically tied to the CF1 PUBSYS L backplane signal, which is pulled high if not actively driven low on the bus, and when the M8191 KTJ11-B is present in a UNIBUS system that will actively drive the CF1 PUBSYS L backplane signal low.

(If I understand the 2.11BSD boot code correctly it checks for the presence of a UNIBUS Map Register at 17770200 instead of checking the Maintenance Register UNIBUS System bit to distinguish between what it considers to be an 11/83 or 11/84, and between an 11/93 or 11/94. I don't know what DEC operating systems might do).

 

[Hunta]

Formally, Unibus registers are needed for controllers that use only a 18-bit address in a 22-bit system, but in fact, at least as far as I know, such controllers are designed for the Unibus bus. Although Unibus was usually 18-bit, there was a variant of 22-bit Unibus - Extended Unibus or EUB. Support for this Unibus was like only in two machines - PDP-11/24 and PDP-11/44. And it seems that only memory was produced as EUB devices. But, besides, there were Unibus controllers that used a 22-bit address for memory addressing - bypassing the Unibus. For example, RH70.

The RESORC program in RT-11 distinguishes between 83-84 and 93-94 systems by a bit in the register. If I remember it right. For the RSX, it need to look at the source code.