Info needed: HP1000F/2117F
Since I have two 2117F's at home, I'll answer this one...
You wrote...
>And one of these cases where documentation can be TOO old... My one
>vintage HP catalog predates these CPUs by a couple of years. The catalog
>only lists the HP2116.
Your machines are F series, the very last of the 1000 line I believe. The
2116 is one of the first. You've got both ends of the spectrum there ;)
Bear this in mind, memory related cards go in the front card cage, i/o
related cards in the rear card cage. There is no relationship between slot1
in front and slot1 in back. In the front card cage, the top three slots must
be used for a specific card in each slot, in the remaining slots order isn't
important (there are a few exceptions to this). In the rear, order isn't
important except the slot number (octal select code) determines the
interrupt priority of the board. Some OS's in the 1000 line expect certain
cards to be in certain slots. In addition, the OS on your disks was likely
genned to expect a certain card in every slot. I'm at work at the moment, so
if there's any of the following notes I'm not sure of, email me over the
weekend when I'm at home and I can clarify.
> DCPC
Dual Channel Port Controller this is basically a card that provides two
independent DMA channels for any I/o cards in the rear panel. Must go in top
slot, front card cage. Attach ribbon cable.
> 13037 intf
The 13037 is a disk controller for 7900A, 7905, 7906, 7920, and possibly
7925 drives. Different models of the 13037 could also support HP-IB as well
as direct connect style drives. This card should go to a 5 inch tall rack
mount box which contains a microprocessor board and several interface
boards. Cables from the 13037 rack mount box then go to the drives.
> Memory Protect
This is an optional memory protect card. ISTR it has to do with handling
parity errors instead of just halting the cpu. This must go in the front
cage, second slot from the top.
> Grd true in/out
Grd true in/out is an I/o card used for line level interfacing (process
control, etc.)
> M.E.M.
This is a Memory Expansion Module. It was necessary if the machine contained
more that 64k of ram. Bear in mind that it is only usefull if the ROM's that
go with it are in the Firmware accessory board. The firmware accessory board
usually attached to the bottom cpu card and hung underneath it on the back
right. The firmware roms provided some extra instructions for memory
management (ie. setting up paging). This must go in the front cage, third
slot from the top.
> Grd true in/out
Same as above.
> 64k HSM
64k ram card.
> 12747H Microcircuit
The microcircuit interface ISTR is a 40 bit card to interface to external
TTL levels. I'll have to look this up. Might also have been used to connect
an I/O extender box, don't recall at the moment.
> 64k HSM
64k ram card.
>12747H BACI 12966A
The 12966A baci card (buffered async control interface) is an RS232 port for
you. One of the better models, actually.
> 64k HSM 12747H
another 64k ram card.
> 7970 Mag Tape 2
The 7970 mag tape 2 card isn't for a second tape drive, it's the 2nd card in
a two controller set for the 7970 tape drive.
> 64k HSM 12747H
another 64k ram card.
> 7970 Mag Tape 1
The 7970 mag tape 1 is part of a 2 card set for a 7970 1/2 tape drive. the
real number is 13181 or 13183 depending on the model 7970 it went to.
> 64k HSM
another 64k ram card.
> Line Printer
Dunno about the line printer designation, but it should be obvious. probably
gpib interface.
> 64k HSM 12747H
another 64k ram card.
> Time Base Gen
The time base generator is a card used to provide various timing circuits.
It does more than just act as a system clock, but that's the best way to
describe it. Whether a TBG card is needed depends on the OS you load and the
software features (modules) you use with it.
> Mem Contr 2102E F.E.M.
This is the main memory controller for all those 64k ram cards. there is a
ribbon cable to attach it to each 64k ram card.
> Main Logic? (under chassis)
This is the cpu card itself. Look for the firmware accessory board attached
underneath the cpu card in the back right. It holds instruction set
additions typically.
>Machine 2 Front card cage Rear card cage
>
> DCPC
same as above
> Jumper
This was used in the I/o card cage if a card was not present. It just
preserved the interrupt priority chain; normally empty non-adjacent slots
are not allowed (unless you're doing all polling mode without need for
arming the interrupts).
> Memory Protect
> I/O Buffer
> M.E.M.
all same as above
> 8 Chan Mux
I would need a part number, but sounds similar to the 12920/12921 mux
controller. Multiple serial ports for terminals, printers, etc.
> Standard Memories Bus I/O
> (256k memory)
> Standard Memories
> (256k memory)
Not familiar with these, they sound 3rd partyish to me, probably memory...
> Disc Intf 2
the second controller of a two controller set, likely a 13210. This was used
ONLY for 7900A disk drives.
> 256kw 12749H
256kw ram.
> Disc Intf 1
the first controller of a two controller set, likely a 13210. only 7900a
again.
> 256kw 12749H
256kw ram
> Mem Contr 2102E
same as above system
> BACI 12966H
same as above system
> 12821A Disc Intf
I don't recall for sure, but I *THINK* this might be for fixed head disks
(earlier than the 7900A's). 2313 I think? Ahhh.. bad memory in my brain :)
> Time Base Gen
same as above
> F.E.M.
FEM or MEM? in a card cage, I would expect it to be MEM, same as above.
Otherwise, an educated guess would be some type of firmware board.
> Main Logic? (under chassis)
same as above.
>1, 2, 3, 4
Need to look up the part number for those. Possibly PSI's (programmable
serial interface), used to hook up graphics terminals (one DB25 for
keyboard, one DB25 for display).
>I still need to find docs on this critter and the cards.
I have a virtually complete documentation set for the above, sans a few of
the cards.
>After the usual pre-launch checks, all of the (apparently) optional cards
>were removed from the card cages and the units were powered up. Curiously,
>they both act identically in that they seem to have some front panel
>function, but the CPUs seem to be hung pretty hard.
The front panel on the 21mx line is pretty straigforward. A&B are registers,
M is the memory location you want to look at, T is the contents of that
location, P is the program counter, and S is a general purpose status
register. The only thing not intuitive is that only when the T register is
displayed and you press the store switch, the M register is automatically
incremented for you (any other time you use the INC M button). Initially I
found this annoying, but it is handy. Note that the A and B registers are
memory locations 0 and 1 respectively - so they should be the same. As a
test, try clearing register a, set a bit pattern, then press store. Then set
the M register to zero, store, then look at T. It should be the same pattern
you stored in A.
Other important notes for a checkout - upon powerup, look at the very bottom
row of lights (a/b/m/t/p/s). Only one should be lit. If more of them are
lit, it indicates several different problems, the most likely of which is
memory configuration/parity error. If the machine has the power fail option,
deadness upon powerup can mean that the batteries are no longer sustaining
memory. The system disables some functions to alert you of this. There is a
way to clear this condition, but I need to check the manual when I get home.
Another thing you can try is selecting the S register, clearing all bits,
store, then set bits 6 through 11 to octal 13 (doesn't matter if there's a
card in slot 13). Press preset, then IBL. If the overflow light comes on,
nothing was loaded from ROM and you have a real problem. Otherwise... Then
point the M register to 37777 and look at the T register. Instead of the
zeros normally present on power up, you should see the contents of boot rom
loader 00 are present (hit INC M to step up). The changing pattern will tell
you it was able to load the boot loader. The 37777 memory location moves
depending on how much ram is in the system.
Finally, the 2117F should go through a powerup POST type test. Depending on
battery charge (if the power fail option is present) sometimes it can take
20 seconds sometimes 30 minutes or so, but eventually if the system is 100%
you should see the leftmost 5 bits or so on the display counting up as it
checks memory.
Since I'm at work and not at home, I'm reciting all this from memory. I
could have easily left something out or said something wrong. Feel free to
email me over the weekend on anything you have questions about.
>No odd sounds or loss of magic smoke, so an initial suspicion is a
>configuration error common to both units. The card cages only have
>specific card notations on a couple of slots, so there is the obvious
>question of proper card positioning. (no idea if the cages are a parallel
>bus or not)
The machines in question are using HSM type memory, which had a slightly
strange setup compared to the memory subsystems in the other HP systems I
also have (2113B 2109B) and work with more often. There are some real
specifics related to slots and ribbon cables in the front card cage. When I
get home I'll check how my 2117F's (which are running fine) are configured.
Jay West
Received on Fri May 28 1999 - 13:19:48 BST
This archive was generated by hypermail 2.3.0
: Fri Oct 10 2014 - 23:32:27 BST