HP9845 service toolkit

[Tony Duell]

> 
> Tony wrote:
> > My workbench is currently covered with a 9845 is many bits :-). And I am
> > slowly figuring out just what's inside that darn machine...
> 
> I don't know about the service toolkit.
> 
> There were apparently major software changes between the 9845A and 9845B.

I believe mine is a 9845B. I've not got the case parts where I can easily 
see them (the PCBs being more interesting are what cover my bench :-)), 
but some of tha add-on ROM modules say '9845B' on them. I have a 
monochrome monitor.

Incidentally, one of the ROM modules is a 3rd party one. Rather than 
being an HP ceramic hybrid in a plasic holder, this is a little PCB with 
2 2764 EPROMs on one side, a couple of '373s and a small PROM (address 
decoder, I guess) on the other. This will be useful in figuring out 
pinouts, etc.

> 
> However, for the 9845B and 9845C there are two different processor
> options.  The original uses two HP "BPC" microprocessors, as described
> in HP Journal articles.  This is the same 16-bit microprocessor used
> in the 9825 and 9835, and in various peripherals including the 9872

I haev one of those. I don't think it's the same device -- exactly -- as 
in a 9825, it appears to have more connections for a start. It is almost 
certainly the same architecture, though.

> plotters.  Later there was a higher-performance CPU option using
> commercial bit-slice parts.  Probably 2901s, but I've never seen one

I have that. It's 3 PCBs with a little backplane on top to link them 
together. One of the PCBs goes into the main backplane, and seems to 
contain mostly TTL, with 8 of what appear to be RAMs (but I can't 
identify them exaclty yet. All have the same 1816-xxxx number, all are 
made by AMD). The middle board ofthe 'sandwich' contains 4 2901 ALUs (you 
are right) + more TTL. The last board has a 2910 sequencer, some ROMs 
(microcode store, I guess), more TTL, including a 74150 which I assume is 
the con=dition bit multiplexer. 

> so I can't say with any certainty.  I'd expect that the I/O processor
> would still be a BPC.

It is, or at least it is in my machine. I've traced the I/O bus 
connections to that board anyway -- the one thing I know for sure is the 
pinout of the I/O slot. Incidentally, the I/O backplane contains a lot of 
buffer chips, a bit like a cut down 9878. 

-tony